Why the Environment Beats the Lesson Plan
In this lesson: Byron makes the case that a strong learning environment — doing, feedback, a tribe, and LLMs — drives roughly 90% of an engineer’s growth, while the formal curriculum contributes only 10%.
Get reminders
One signup gets you join links, reminders, and the recording library. No ongoing commitment.
Top 3 takeaways
Environment beats the lesson plan
The 70-20-10 model shows formal training is just 10%; on-the-job doing and learning from others drive the other 90%.
Doing plus feedback compounds learning
Building beats consuming, and fast feedback cements it — like violinists’ 10,000 hours of focused, feedback-driven practice, not passive repetition.
Your tribe determines whether you finish
Cohort-based programs see 87–90% completion versus 3–15% for solo learning; shared struggle keeps people moving instead of quitting.
Byron Mackay
Director of Learning, Gauntlet AI
Byron Mackay is Director of Learning at Gauntlet AI, training hundreds of engineers to work AI-first. Previously a Principal AI Engineer, he led curriculum development at BloomTech.
Lesson notes
A written walkthrough of the lecture, covering the model, the practice, and the things that trip people up.
The 70-20-10 model
Byron opens with the Center for Creative Leadership's 70-20-10 model: roughly 70% of effective learning comes from on-the-job experience, 20% from interacting with others, and only 10% from formal training. His point as a professional educator: formal instruction is necessary but small, and the real acceleration comes from combining all three rather than living in a build-only silo.
Doing beats consuming
Reading books and watching videos produce ideas and direction but don't make you an expert. Doing does — and it applies to LLMs too. Gauntlet leans on coding agents but pushes engineers to do more with them (timed mini-hackathons, longer planning sessions) so they build new skills rather than just shipping for its own sake.
Doing + feedback
The strongest lever is doing plus feedback. Byron cites the violinist studies: the top performers weren't separated by raw hours but by 10,000 hours of focused, feedback-driven practice. This is why Gauntlet emphasizes "building in public," runs 3–5 feedback-generating deliverables per week, and prizes fast correction loops.
The tribe
Cohort-based programs complete at 87–90% versus 3–15% for self-directed learning (including solo Coursera-style courses). The people around you turn struggle into a shared problem rather than a reason to quit. Gauntlet's "tribe" includes cohort mates, instructors and office hours, engineering experts, and hiring partners — but tribes can also be built via Slack groups, Reddit, meetups, and networking.
Sharing ideas: architecture defense
A recent Gauntlet practice: five people get the same case study, each designs a solution within hours, then presents for five minutes while peers spend five minutes poking holes. It has produced a remarkable difference in the direction people take afterward.
Learning with LLMs
LLMs are a powerful but partial piece of the process. Byron references a fall-2023 Harvard physical-science study where an AI tutor (heavily grounded to avoid hallucination, with a control group that alternated weekly) produced higher engagement, better grades, and stronger explanations. Because LLMs converge on the mean, they get you to average fast — then you springboard. He notes scaffolded learning (quizzes, case studies), a system-design skill he built, and Gauntlet's internal scaffolded AI tutor, "Traverse." The TypeScript "discriminated unions" example illustrates why breadth of knowledge still matters: LLMs won't reach for lesser-known tools unless you know to ask.
The role of curriculum (the 10%)
Curriculum is the smallest slice but still essential — it provides focus, direction, and checkpoints so learners don't stall deciding what's next. Put a strong environment on top of a lesson plan and, in Byron's framing, there's no stopping you.
Q&A highlights
On agentic workflows: context is king. Rather than chasing frameworks, invest in well-placed CLAUDE.md (or AGENTS.md) context files across your codebase, build specialized agents per area (frontend, backend/database), then add an orchestrator once task-giving exceeds what you can manage. Recommended repo: awesome-skills; a favorite is the "grill me" skill for pressure-testing ideas and PRs. The eligibility and assessment requirements discussed in this recording reflect guidance at the time. Current requirements differ by track: Gauntlet for America requires US citizenship; Gauntlet Prime requires the ability to work in the US without employer sponsorship. Admissions reviews assessment results alongside engineering experience and the rest of the profile. Graduates typically land AI-focused engineering roles via hiring partners after an in-person Austin interview.
Transcript
Read transcript
Why the Environment Beats the Lesson Plan
00:00:04.960Ash Tilawat: Welcome to Night School. I'm Ash, the
00:00:07.040Ash Tilawat: CTO of Gauntlet AI. Over the last 2
00:00:09.679Ash Tilawat: years, Gauntlet has trained over,200
00:00:11.679Ash Tilawat: engineers on how to code with AI, use
00:00:14.480Ash Tilawat: coding agents, build harnesses, and
00:00:16.320Ash Tilawat: software factories, and to build AI
00:00:18.720Ash Tilawat: applications like rag pipelines, agentic
00:00:21.279Ash Tilawat: rag, agents, and graphs. Night School is
00:00:24.640Ash Tilawat: our open opportunity for you to get free
00:00:27.199Ash Tilawat: lessons on all of these topics every
00:00:29.359Ash Tilawat: single week. Each of the teachers are
00:00:32.079Ash Tilawat: members of my staff. They're
00:00:33.840Ash Tilawat: practitioners by day and they teach
00:00:35.760Ash Tilawat: about 7 to 10 hours per week by night.
00:00:38.399Ash Tilawat: We want to give you the hands-on lessons
00:00:40.719Ash Tilawat: that we're giving in our corporate
00:00:42.239Ash Tilawat: training, our gauntlet, and our Gauntlet
00:00:44.399Ash Tilawat: for America programs.
00:00:46.559Ash Tilawat: Please look at the description if you
00:00:48.559Ash Tilawat: want the slides, the repositories, or
00:00:50.879Ash Tilawat: any other materials that are covered in
00:00:52.480Ash Tilawat: class. Please visit us at
00:00:55.199Ash Tilawat: gauntletai.com/nightchool
00:00:57.440Ash Tilawat: if you want to sign up and see the
00:00:58.719Ash Tilawat: topics for the next session. If you join
00:01:00.719Ash Tilawat: us live, you get to ask questions and
00:01:03.120Ash Tilawat: get to be a part of our community.
00:01:05.040Ash Tilawat: Thanks for watching.
00:01:06.080Byron Mackay: Hey there everybody.
00:01:08.720Byron Mackay: Welcome to tonight's night school. My
00:01:11.680Byron Mackay: name is Byron. It's good to see you all.
00:01:15.040Byron Mackay: Well, good to be with you and I'm
00:01:17.200Byron Mackay: excited to talk to you about tonight's
00:01:18.640Byron Mackay: topic. Uh it's it's one that I spend a
00:01:23.119Byron Mackay: lot of time with. I'm the director of
00:01:25.040Byron Mackay: learning here at uh Gauntlet and we're
00:01:27.040Byron Mackay: going to be talking about the
00:01:29.040Byron Mackay: environment, the learning environment,
00:01:31.360Byron Mackay: which uh may not sound like most
00:01:33.840Byron Mackay: exciting, but I'm going to do my very
00:01:35.040Byron Mackay: best to make it exciting. Uh there's a
00:01:38.960Byron Mackay: lot that we do here at Gauntlet that
00:01:41.280Byron Mackay: really focuses on the environment and uh
00:01:45.439Byron Mackay: it's part big part of my job is to make
00:01:48.399Byron Mackay: sure that what we teach and how we learn
00:01:51.920Byron Mackay: and the experience we have here uh helps
00:01:54.799Byron Mackay: you become the best engineer that you
00:01:56.720Byron Mackay: can. Um, especially in the age of AI
00:02:00.640Byron Mackay: when everything's changing and moving so
00:02:02.640Byron Mackay: quickly, it's uh it's it's a hard it's a
00:02:06.479Byron Mackay: it's a it's a hard and interesting
00:02:08.560Byron Mackay: challenge for us to make these
00:02:09.759Byron Mackay: adjustments and the faster we can do it,
00:02:12.640Byron Mackay: the more successful we will be. Um I
00:02:15.120Byron Mackay: talk to a lot of great engineers all the
00:02:16.879Byron Mackay: time and it's so interesting to hear uh
00:02:19.760Byron Mackay: especially ones that have gone through
00:02:21.120Byron Mackay: this before when it was now things are
00:02:24.400Byron Mackay: moved to the cloud now we have mobile
00:02:26.080Byron Mackay: development and now we have AI and then
00:02:30.000Byron Mackay: it all comes back to well now people
00:02:32.319Byron Mackay: still need cobalt developers and these
00:02:34.640Byron Mackay: uh this vast experience of going through
00:02:37.040Byron Mackay: these changes and seeing how things can
00:02:38.560Byron Mackay: revert and how they learn and stayed up
00:02:40.720Byron Mackay: to date um is really inspiring and it's
00:02:43.840Byron Mackay: shaped shaped a little. It's kind of
00:02:45.120Byron Mackay: molded and helped molded what we've um
00:02:47.680Byron Mackay: what we've been able to accomplish here.
00:02:49.040Byron Mackay: So, I'm going to share some of those
00:02:50.080Byron Mackay: learnings with you all today uh as you
00:02:53.440Byron Mackay: all embark and really uh you know
00:02:56.879Byron Mackay: traverse this uh this new age that we're
00:03:00.560Byron Mackay: in. So, without further ado, I'm going
00:03:04.480Byron Mackay: to uh put up some slides and we'll go
00:03:07.360Byron Mackay: through we'll go through this and I'll
00:03:08.959Byron Mackay: talk through it all. Um, and please feel
00:03:11.440Byron Mackay: free uh to ask questions. Um, I will be
00:03:17.040Byron Mackay: uh I will be answering them as we go.
00:03:20.560Byron Mackay: Um, and I have a lot to say. So, please
00:03:22.720Byron Mackay: answer ask ask any questions you have.
00:03:24.640Byron Mackay: Help me uh address you what you want to
00:03:26.720Byron Mackay: learn out of tonight. Um, so that we can
00:03:29.840Byron Mackay: you can get the most out of it. All
00:03:31.599Byron Mackay: right. Give me just a moment here as I
00:03:34.000Byron Mackay: uh pull these up really quick.
00:03:38.799Byron Mackay: Okay.
00:03:41.599Byron Mackay: Let me just get configure my screens
00:03:45.120Byron Mackay: here a bit so I make sure I can see
00:03:46.560Byron Mackay: everything and everybody. Okay.
00:03:49.360Byron Mackay: So,
00:03:52.319Byron Mackay: tonight's presentation is titled why the
00:03:54.560Byron Mackay: environment beats the lesson plan. Uh
00:03:58.080Byron Mackay: there are lots of ways to learn, but
00:04:00.640Byron Mackay: there are also ways to learn fast and
00:04:02.879Byron Mackay: effectively.
00:04:04.480Byron Mackay: And it does not require allnighters. It
00:04:07.519Byron Mackay: does not require your head being in a
00:04:09.040Byron Mackay: book all day and it does not require you
00:04:12.480Byron Mackay: to hit your head constantly against a
00:04:14.239Byron Mackay: problem until you figure it out.
00:04:15.920Byron Mackay: Sometimes we romanticize that. We see
00:04:19.280Byron Mackay: these montages of people learning and
00:04:21.919Byron Mackay: growing and doing all the hard things uh
00:04:24.560Byron Mackay: by themselves. And quite frankly, that
00:04:28.479Byron Mackay: is a one way to learn it. And you'll get
00:04:30.800Byron Mackay: there eventually, but you won't get
00:04:32.160Byron Mackay: there very fast. And so I want to talk
00:04:34.560Byron Mackay: about tonight are some strategies and
00:04:36.479Byron Mackay: some ideas and concepts that you can
00:04:38.240Byron Mackay: take to improve your environment and
00:04:40.720Byron Mackay: what that requires.
00:04:43.280Byron Mackay: So first I want to show this. This is
00:04:46.400Byron Mackay: the 702010 model. The 702010 model was
00:04:50.800Byron Mackay: developed by the center for creative
00:04:52.160Byron Mackay: leadership and it shows you what su what
00:04:54.800Byron Mackay: a successful uh distribution of of
00:04:58.320Byron Mackay: exercises looks like in terms of
00:04:59.919Byron Mackay: learning. Now, I want you to look at the
00:05:03.680Byron Mackay: right side. That is formal training.
00:05:07.440Byron Mackay: That you might think, Byron, that's what
00:05:10.000Byron Mackay: you do for a living is formal training.
00:05:11.840Byron Mackay: You're right. I do. I'm 10% of the
00:05:14.639Byron Mackay: equation. I'm necessary. Well, formal
00:05:17.680Byron Mackay: training is necessary, right? But it's
00:05:20.160Byron Mackay: just a small piece of it. Look at these
00:05:21.919Byron Mackay: other two. Interactions with others,
00:05:24.000Byron Mackay: 20%. It's so important, and we're going
00:05:26.320Byron Mackay: to talk about that. on the job
00:05:28.160Byron Mackay: experience critical, right? Huge piece
00:05:32.000Byron Mackay: of how we learn. So, you combine all
00:05:35.440Byron Mackay: these things together and then you start
00:05:38.160Byron Mackay: to find there's this sweet spot. There's
00:05:40.400Byron Mackay: this place where you can accelerate and
00:05:42.880Byron Mackay: go really, really fast. Um, somebody
00:05:46.479Byron Mackay: mentioned that the slide isn't showing.
00:05:48.000Byron Mackay: I just want to make sure that it is
00:05:50.400Byron Mackay: working. Can anybody else uh same issue?
00:05:53.600Byron Mackay: Okay, so we've got Oh, you know what? I
00:05:56.080Byron Mackay: know why. Hold on. Hold on.
00:06:00.240Byron Mackay: That's the one. Personal problem. User
00:06:03.120Byron Mackay: error. I apologize. So, let me touch on
00:06:06.400Byron Mackay: this one more time real quick. Um,
00:06:09.440Byron Mackay: formal training 10%. Right? Very, very
00:06:12.880Byron Mackay: small piece. Look at this. The other 90%
00:06:15.919Byron Mackay: big chunk of it on the job experience.
00:06:18.160Byron Mackay: Another chunk of it interacting with
00:06:19.680Byron Mackay: others. You put all these together and
00:06:22.800Byron Mackay: now you're going to get ahead. you're
00:06:25.600Byron Mackay: going to move so much faster. Um, I have
00:06:29.120Byron Mackay: been in the boat where I have just been
00:06:31.120Byron Mackay: trying to do the 70% the on the job
00:06:33.360Byron Mackay: experience, right? I've been in that I
00:06:35.680Byron Mackay: would stay there and I would just work
00:06:37.039Byron Mackay: my head down typing away on the
00:06:39.600Byron Mackay: computer, making apps thinking, "Oh, I'm
00:06:42.400Byron Mackay: going I'm learning so much." And to my
00:06:44.960Byron Mackay: to my effort, I was right. But the
00:06:47.919Byron Mackay: problem is is I lived in a silo. I don't
00:06:50.800Byron Mackay: know if any of you have felt that way,
00:06:53.039Byron Mackay: but you live in your own little silo and
00:06:55.039Byron Mackay: you just think, I'm doing it because I'm
00:06:57.199Byron Mackay: building and doing something. It feels
00:06:59.280Byron Mackay: really productive. But at the end of the
00:07:01.520Byron Mackay: day, if you ever poke your head out of
00:07:02.800Byron Mackay: the silo and you look around, other
00:07:04.800Byron Mackay: people are going in different
00:07:05.919Byron Mackay: directions. They've got different ideas.
00:07:08.240Byron Mackay: And it's the ones that interact with
00:07:09.919Byron Mackay: each other that really uh start to
00:07:13.840Byron Mackay: explore and get out of their comfort
00:07:15.759Byron Mackay: zone, start doing more. They get out of
00:07:17.440Byron Mackay: their silos. they get out of their echo
00:07:19.039Byron Mackay: chambers and they start to accelerate
00:07:22.720Byron Mackay: even more. And then there's this piece
00:07:24.800Byron Mackay: of it where we still need some formal
00:07:26.960Byron Mackay: training. We still need to sit down,
00:07:28.400Byron Mackay: have somebody get up there and say,
00:07:30.080Byron Mackay: "This is how we do it and this is the
00:07:33.120Byron Mackay: way it's going." And we put all that
00:07:35.680Byron Mackay: together and suddenly we can move a
00:07:37.599Byron Mackay: whole lot faster. So it's a balancing
00:07:40.319Byron Mackay: act of of a sort that we need to go for.
00:07:43.840Byron Mackay: I want to make that that point very
00:07:45.840Byron Mackay: clear because I think it's very easy as
00:07:47.199Byron Mackay: engineers and especially in the world
00:07:49.360Byron Mackay: where we have an LLM that's constantly
00:07:51.360Byron Mackay: on that we can just talk to all the
00:07:52.880Byron Mackay: time. It doesn't count that doesn't
00:07:55.120Byron Mackay: count as as your friend as you're
00:07:56.800Byron Mackay: interacting with your other I'm just
00:07:57.919Byron Mackay: going to say that out loud right now.
00:07:59.360Byron Mackay: They have a place they're very important
00:08:01.199Byron Mackay: very powerful but they're not your
00:08:03.440Byron Mackay: friend that you're going to talk to. So
00:08:05.120Byron Mackay: we'll we'll talk about a little more of
00:08:06.560Byron Mackay: that in a second here. But I first want
00:08:08.879Byron Mackay: to focus on that 70%.
00:08:12.400Byron Mackay: doing is better is more than consuming.
00:08:15.599Byron Mackay: Right? I think that we I think we can
00:08:18.800Byron Mackay: all attest to this in our careers that
00:08:20.960Byron Mackay: reading a book doesn't make you an
00:08:22.400Byron Mackay: expert. Right? Listening to YouTube
00:08:25.360Byron Mackay: videos doesn't make you an expert. It
00:08:26.879Byron Mackay: produces some ideas. It gives you some
00:08:29.120Byron Mackay: direction, but it's not going to take
00:08:31.520Byron Mackay: you from where you are to an expert
00:08:34.080Byron Mackay: level. There needs to be doing. and
00:08:36.640Byron Mackay: doing is going to take you far further
00:08:39.200Byron Mackay: than consuming ever will because there's
00:08:41.599Byron Mackay: just things that you will not learn just
00:08:43.919Byron Mackay: from consuming. Now I'm this is also
00:08:47.760Byron Mackay: true when it comes to learning with
00:08:49.760Byron Mackay: LLMs. LM can get us over humps and
00:08:52.000Byron Mackay: hurdles and do all sorts of great things
00:08:53.519Byron Mackay: for us but we still have to do it. And I
00:08:56.080Byron Mackay: worry sometimes that as engineers uh
00:09:00.320Byron Mackay: with coding agents as good as they are
00:09:03.200Byron Mackay: that we're missing that boat. That's one
00:09:05.279Byron Mackay: thing at Gaunt that we really strive for
00:09:07.040Byron Mackay: where we use coding agents, we leverage
00:09:08.880Byron Mackay: coding agents, but we push ourselves to
00:09:11.279Byron Mackay: do more with them. We push ourselves to
00:09:14.000Byron Mackay: do a lot more with them so that we can
00:09:15.680Byron Mackay: build these new skill sets that we need
00:09:17.600Byron Mackay: today. Um, we have very difficult
00:09:20.320Byron Mackay: assignments. Sometimes we do many
00:09:21.839Byron Mackay: hackathons where it's very time
00:09:23.440Byron Mackay: constrained and we're seeing how fast we
00:09:25.680Byron Mackay: can push the agents. Other times we do
00:09:28.560Byron Mackay: longer planning sessions so we can talk
00:09:30.959Byron Mackay: it through. And all of these things are
00:09:34.320Byron Mackay: ways that you can uh you can stop and
00:09:38.399Byron Mackay: have different perspective and really
00:09:41.040Byron Mackay: analyze what we're trying to accomplish
00:09:43.120Byron Mackay: here. Not just building, not just
00:09:44.640Byron Mackay: building for building sake. Um but
00:09:46.959Byron Mackay: rather we're building and we're doing
00:09:49.279Byron Mackay: these these exercises that get our minds
00:09:52.560Byron Mackay: thinking again, that get our minds
00:09:54.000Byron Mackay: really churning and doing. But there's
00:09:56.399Byron Mackay: an there's even more that we can do.
00:09:59.040Byron Mackay: There's one better than doing
00:10:02.399Byron Mackay: and that's doing plus feedback.
00:10:05.920Byron Mackay: Um, this is one one this is one of those
00:10:08.399Byron Mackay: cases where you step out of the echo
00:10:10.160Byron Mackay: chamber and you go in and actually get
00:10:12.399Byron Mackay: this feedback. It's it's amazing how
00:10:15.839Byron Mackay: quickly you can learn. And there's
00:10:17.839Byron Mackay: actually full there's a lot of studies
00:10:19.839Byron Mackay: around this. In fact, there's one uh
00:10:22.560Byron Mackay: where they did it uh on violinists
00:10:26.000Byron Mackay: and they asked they they they kind of
00:10:28.959Byron Mackay: grouped violinists into groups saying
00:10:30.640Byron Mackay: these are the top tier violinists. Um
00:10:32.880Byron Mackay: these are your your mid-tier violinists.
00:10:34.880Byron Mackay: And then they had the third group they
00:10:36.000Byron Mackay: called the the violin teachers. And they
00:10:39.680Byron Mackay: said, "Okay, how much focused time did
00:10:44.079Byron Mackay: you do growing up uh to become this
00:10:48.240Byron Mackay: proficient at uh the violin?" Well, how
00:10:50.800Byron Mackay: many hours did you put in? And uh who
00:10:55.279Byron Mackay: who would who would want to take a
00:10:56.720Byron Mackay: guess? The number is quite a popular
00:10:58.480Byron Mackay: number. I'll just throw that hint out
00:11:00.000Byron Mackay: there. But who can guess how many hours
00:11:02.959Byron Mackay: they put in to be an an expert?
00:11:11.040Byron Mackay: Steven nailed it. Steven nailed it. It's
00:11:14.399Byron Mackay: 10,000 hours, but not just 10,000 hours
00:11:16.880Byron Mackay: of practice. 10,000 hours of focused,
00:11:20.320Byron Mackay: feedbackdriven practice.
00:11:23.200Byron Mackay: And that was the that was the key
00:11:24.640Byron Mackay: differentiator. The ones in the middle
00:11:26.720Byron Mackay: also put in a lot of hours, but they may
00:11:28.240Byron Mackay: not have had the feedback that the that
00:11:30.640Byron Mackay: the better group had. Um, it was a that
00:11:35.600Byron Mackay: feedback loop was was impressively
00:11:39.519Byron Mackay: impactful
00:11:41.279Byron Mackay: and that's something that we want to
00:11:42.880Byron Mackay: make sure that we all get. And so we
00:11:45.440Byron Mackay: always talk about building in public.
00:11:47.040Byron Mackay: We've heard that that's a actually a
00:11:49.120Byron Mackay: really great concept when it comes to
00:11:51.839Byron Mackay: getting better faster. Now, there is a
00:11:54.800Byron Mackay: side of side of you that might like if
00:11:57.120Byron Mackay: you're like me, you might go, "Oh, I'm a
00:11:59.839Byron Mackay: little nervous about this because I
00:12:01.920Byron Mackay: don't want people to say my stuff is
00:12:03.519Byron Mackay: bad." Well,
00:12:06.000Byron Mackay: we might have to just start getting used
00:12:07.279Byron Mackay: to it. If we want to get faster, that's
00:12:10.160Byron Mackay: the only that's one of the few ways
00:12:11.920Byron Mackay: we're actually going to make sure we get
00:12:13.120Byron Mackay: out of ourselves and into a a better a
00:12:16.160Byron Mackay: better workplace is by getting that
00:12:18.240Byron Mackay: feedback from others. Um, Miriam had a
00:12:21.120Byron Mackay: question here. She said, "How do I get
00:12:23.680Byron Mackay: over or not heavily focused on the
00:12:25.760Byron Mackay: consuming part when learning is a huge
00:12:28.800Byron Mackay: part of the process?" That's a great
00:12:31.760Byron Mackay: question and uh one that I've personally
00:12:34.560Byron Mackay: been kind of mulling over myself. Um
00:12:39.920Byron Mackay: the the way we we we do it is one way of
00:12:43.200Byron Mackay: doing is to always make sure we're
00:12:44.560Byron Mackay: building. We still learn we have we set
00:12:47.360Byron Mackay: aside time to consume and consuming is
00:12:50.160Byron Mackay: important. it can expand our thoughts,
00:12:51.680Byron Mackay: expand our ideas. Um, we really as
00:12:54.800Byron Mackay: engineers, we need to know the tools
00:12:56.399Byron Mackay: that are out there. We need to know the
00:12:58.240Byron Mackay: tools that exist that we can leverage,
00:13:01.360Byron Mackay: that we can build with. And to do that,
00:13:03.279Byron Mackay: we have to consume a lot of information,
00:13:05.680Byron Mackay: right? Um, I'll give you an example.
00:13:08.560Byron Mackay: This is one of my favorite examples.
00:13:10.000Byron Mackay: It's it's true every time I I talk about
00:13:12.079Byron Mackay: it with um
00:13:14.959Byron Mackay: we have a lot of a lot of engineers know
00:13:17.040Byron Mackay: Typescript. But I imagine many in our
00:13:18.160Byron Mackay: audience tonight know TypeScript or have
00:13:20.240Byron Mackay: worked with TypeScript. And Typescript
00:13:22.160Byron Mackay: is a great language, very powerful, lot
00:13:24.000Byron Mackay: of power inside Typescript. There's a
00:13:26.639Byron Mackay: feature in Typescript called
00:13:28.720Byron Mackay: discriminated unions.
00:13:31.040Byron Mackay: And discriminated unions whenever I
00:13:33.120Byron Mackay: survey a group of TypeScript engineers.
00:13:35.760Byron Mackay: It's not a topic that that a lot of
00:13:37.600Byron Mackay: people actually know. Some do, but many
00:13:40.560Byron Mackay: don't. Now, if you know discriminated
00:13:43.200Byron Mackay: unions, um they're quite powerful.
00:13:47.040Byron Mackay: If you don't know what they are, they're
00:13:48.240Byron Mackay: they're kind of like interfaces
00:13:50.720Byron Mackay: where you can define a a shape and then
00:13:54.160Byron Mackay: expand that shape, that data shape. Um,
00:13:57.360Byron Mackay: but it's in a very JavaScripty way
00:13:59.279Byron Mackay: that's very great inside of Typescript
00:14:02.720Byron Mackay: because most people don't know what
00:14:04.399Byron Mackay: discriminated unions are or they don't
00:14:05.920Byron Mackay: leverage them. Asking an LLM to build a
00:14:08.959Byron Mackay: TypeScript app, it's probably not going
00:14:10.959Byron Mackay: to use discriminated unions. Why?
00:14:14.320Byron Mackay: Because discriminated unions are not
00:14:16.399Byron Mackay: part of the mean. They're outside of
00:14:18.720Byron Mackay: that. And LLMs converge on the mean.
00:14:22.560Byron Mackay: You're going to get average output,
00:14:25.120Byron Mackay: which sometimes, most of the time, I'd
00:14:27.040Byron Mackay: imagine is probably just fine. But if
00:14:29.279Byron Mackay: there's a case where you're like, "No, a
00:14:30.800Byron Mackay: discriminated union is exactly what we
00:14:32.639Byron Mackay: need right here." You have to be
00:14:34.480Byron Mackay: specific and tell it that, and then
00:14:36.959Byron Mackay: you'll get your discriminated union and
00:14:38.480Byron Mackay: be able to leverage that piece of the of
00:14:40.079Byron Mackay: the of the language.
00:14:42.320Byron Mackay: So I say all that to say we must consume
00:14:45.600Byron Mackay: knowledge because we must be able to
00:14:48.639Byron Mackay: know what tools are available. We must
00:14:50.560Byron Mackay: have the breath. But at the same time,
00:14:52.720Byron Mackay: if we're not building with that
00:14:54.240Byron Mackay: knowledge, it's going to sit dormant and
00:14:56.240Byron Mackay: we're going to forget it. Our brains
00:14:59.199Byron Mackay: will retain more if there's more rigor
00:15:01.440Byron Mackay: to remember and implement and understand
00:15:04.399Byron Mackay: than if it's just than if there's less
00:15:06.320Byron Mackay: rigor. I would say reading a book is
00:15:08.880Byron Mackay: considered less rigor in my opinion than
00:15:11.440Byron Mackay: trying to implement the thing in real
00:15:13.440Byron Mackay: life. When we implement it, that's when
00:15:15.839Byron Mackay: the that's when the rigor sets in and
00:15:17.440Byron Mackay: our brains are really focused on going,
00:15:19.440Byron Mackay: "Oh, this is important stuff. I really
00:15:21.120Byron Mackay: need to remember this." And then when we
00:15:23.360Byron Mackay: go out into the world and we get
00:15:24.959Byron Mackay: feedback on top of that, then we really
00:15:28.320Byron Mackay: get that rigor because that's that's
00:15:29.839Byron Mackay: emotional. That that can get very
00:15:31.440Byron Mackay: emotional for some of us. and we get
00:15:33.600Byron Mackay: that feedback and all of a sudden we're
00:15:36.079Byron Mackay: going, "Oh, this is this is important.
00:15:38.720Byron Mackay: Oh, I see what that person said. They
00:15:40.880Byron Mackay: they criticized me for this over here.
00:15:43.360Byron Mackay: Maybe it's because I didn't use
00:15:44.480Byron Mackay: discriminated union when I should have.
00:15:46.240Byron Mackay: I'm not going to forget that as easily
00:15:47.760Byron Mackay: next time." And now I've just bumped
00:15:50.240Byron Mackay: myself out beyond just the mean and I
00:15:53.440Byron Mackay: can go and I can work a lot better and
00:15:54.880Byron Mackay: be more effective. So,
00:15:58.399Byron Mackay: Miriam, to to circle back to your
00:16:00.639Byron Mackay: question, uh, it's always good to be
00:16:04.320Byron Mackay: building. Um, always have a project that
00:16:07.279Byron Mackay: you're trying to implement, uh, that
00:16:09.440Byron Mackay: you're always working on. And it doesn't
00:16:11.040Byron Mackay: have to be something that you're reading
00:16:12.560Byron Mackay: about. I'm not saying you shouldn't you,
00:16:15.279Byron Mackay: uh, what I you can be reading about
00:16:16.800Byron Mackay: something totally t you know, totally
00:16:18.800Byron Mackay: separate from what you're working on.
00:16:20.720Byron Mackay: It's still going to come into play. Just
00:16:22.399Byron Mackay: having those thoughts will come into
00:16:24.160Byron Mackay: play in the building. you'll have a
00:16:25.600Byron Mackay: chance to kind of push things on and get
00:16:27.920Byron Mackay: better there. So,
00:16:31.040Byron Mackay: all right. So, doing feedback big pieces
00:16:34.399Byron Mackay: and at at Gauntlet, we really push this
00:16:37.360Byron Mackay: idea a lot. We want you to get lots of
00:16:39.199Byron Mackay: feedback. And so, we have multiple
00:16:41.440Byron Mackay: deliverables every week. Uh we're up we
00:16:44.639Byron Mackay: we have at least three we get up to
00:16:47.040Byron Mackay: four, sometimes five different
00:16:49.120Byron Mackay: deliverables every week where we're
00:16:50.560Byron Mackay: giving you feedback or you're getting
00:16:52.079Byron Mackay: feedback in one way or another. Um, and
00:16:54.800Byron Mackay: that's really important so that you can
00:16:56.639Byron Mackay: see where you're going and what you're
00:16:57.920Byron Mackay: doing well and what you can improve on.
00:17:00.880Byron Mackay: Um, and if you can build where someone
00:17:03.600Byron Mackay: can correct you fast, that is even more
00:17:06.400Byron Mackay: critical. The sooner the feedback, the
00:17:09.439Byron Mackay: the more you're likely you're going to
00:17:11.199Byron Mackay: remember it and implement it and move on
00:17:13.039Byron Mackay: faster.
00:17:14.880Byron Mackay: So, you want to create that environment
00:17:16.160Byron Mackay: for yourself. Um, you can do it here at
00:17:18.959Byron Mackay: Gauntlet. There's ways of doing it your
00:17:20.959Byron Mackay: own outside of Gauntlet, but you want to
00:17:22.559Byron Mackay: make sure you can get that feedback loop
00:17:24.240Byron Mackay: pretty quick.
00:17:26.799Byron Mackay: All right,
00:17:29.039Byron Mackay: next piece. The tribe.
00:17:32.080Byron Mackay: The tribe is so important. Um, and I
00:17:36.480Byron Mackay: have a lot of personal experience with
00:17:38.320Byron Mackay: this one because I've tried both uh both
00:17:41.280Byron Mackay: learning formats here when it comes to
00:17:44.320Byron Mackay: um formal trainings. Um,
00:17:48.720Byron Mackay: so look at these look at these numbers
00:17:50.320Byron Mackay: here.
00:17:51.919Byron Mackay: If you are in a cohort-based program,
00:17:54.320Byron Mackay: you have an 87 to 90% chance of
00:17:57.440Byron Mackay: completing it as opposed to a
00:18:00.160Byron Mackay: self-directed
00:18:01.840Byron Mackay: learning, which is 3 to 15%. By the way,
00:18:04.559Byron Mackay: that's also true for when I say
00:18:06.960Byron Mackay: self-directed, I also include in there
00:18:09.919Byron Mackay: um things like Corsera courses, things
00:18:12.880Byron Mackay: where you're just doing it by yourself,
00:18:14.720Byron Mackay: where you're into, you're just learning
00:18:16.160Byron Mackay: things alone. Um, and I have implemented
00:18:20.960Byron Mackay: both programs in my career multiple
00:18:22.960Byron Mackay: times, and this holds this holds very
00:18:27.120Byron Mackay: accurately to my experience. um that a
00:18:30.640Byron Mackay: cohort-based program you are far more
00:18:32.559Byron Mackay: likely to complete it than if you're
00:18:35.360Byron Mackay: trying to do it by yourself. Um
00:18:38.480Byron Mackay: and a big part of that is because of the
00:18:40.640Byron Mackay: people you're around, right? If you are
00:18:43.919Byron Mackay: sitting in a college class, you're more
00:18:46.320Byron Mackay: likely to stick through it rather than
00:18:48.480Byron Mackay: watching a video on your computer. Um
00:18:51.600Byron Mackay: it's it's also true when you are um
00:18:55.919Byron Mackay: working at a company, right? you have
00:18:58.080Byron Mackay: your team. Uh why do people have
00:19:00.559Byron Mackay: co-founders? Because of this.
00:19:04.320Byron Mackay: Because if we go try to go in on
00:19:06.080Byron Mackay: ourselves, we're pro we're far less
00:19:07.919Byron Mackay: likely to be successful than if we have
00:19:09.440Byron Mackay: people who are there to help us build
00:19:11.440Byron Mackay: and grow, right? This line here at the
00:19:13.600Byron Mackay: bottom is is I think very poignant and
00:19:16.240Byron Mackay: and and sings a little bit to me. But
00:19:19.039Byron Mackay: it's struggles to become uh struggles
00:19:22.320Byron Mackay: become the shared problem instead of a
00:19:25.280Byron Mackay: reason to quit. And I think that's so
00:19:27.679Byron Mackay: true in my experience, as I'm sure
00:19:29.840Byron Mackay: others can relate. But yes, the
00:19:31.840Byron Mackay: struggles we have when we're learning on
00:19:33.840Byron Mackay: our own can sometimes stop us and go
00:19:35.520Byron Mackay: like, I don't even want to do it. Um, I
00:19:38.320Byron Mackay: can tell you right now, I'm learning all
00:19:40.000Byron Mackay: about security. All about security. I
00:19:42.400Byron Mackay: have all these uh I'm learning so much
00:19:44.480Byron Mackay: as I'm learning about AI and security
00:19:46.799Byron Mackay: and it's a little bit hard for me to
00:19:48.880Byron Mackay: kind of like to to want to go and build
00:19:52.240Byron Mackay: and try to really get into the the weeds
00:19:55.600Byron Mackay: of security. I need somebody else to
00:19:58.080Byron Mackay: come with me.
00:20:00.320Byron Mackay: And so I'm working on that. So I'm
00:20:02.160Byron Mackay: working on on getting people to come
00:20:03.760Byron Mackay: with me on that journey so that I can I
00:20:06.000Byron Mackay: can have I can share these problems and
00:20:07.679Byron Mackay: we can get over them together. Um, here
00:20:10.559Byron Mackay: at Gauntlet, we have a tribe of a uh
00:20:14.480Byron Mackay: that consists of your cohort mates. Uh,
00:20:17.120Byron Mackay: that's a big part of what we do, a big
00:20:19.360Byron Mackay: reason why Gauntlet is so impactful, I
00:20:22.480Byron Mackay: think, for people. It's not just what
00:20:23.840Byron Mackay: you learn, but the network you're
00:20:25.280Byron Mackay: building while you're here and your
00:20:26.400Byron Mackay: cohort, the people that are in there
00:20:28.080Byron Mackay: with you that are doing the same things
00:20:29.440Byron Mackay: with you is a big piece of that. You
00:20:31.440Byron Mackay: have, of course, your instructors that
00:20:33.679Byron Mackay: are teaching you. Um, every week we have
00:20:36.640Byron Mackay: a number of lectures and these
00:20:38.000Byron Mackay: instructors come in and they all have
00:20:39.600Byron Mackay: their different flavor and different
00:20:40.720Byron Mackay: style of doing things. Um, but they're
00:20:42.400Byron Mackay: there to help you and they we hold
00:20:43.760Byron Mackay: office hours to for those instructors to
00:20:46.159Byron Mackay: ask to help you with your questions and
00:20:48.000Byron Mackay: your projects. We have a great team of
00:20:50.400Byron Mackay: engineering experts. Um, and these
00:20:53.039Byron Mackay: engineering experts uh are really above
00:20:55.440Byron Mackay: and beyond. They are building so much
00:20:58.080Byron Mackay: and they uh they come and teach and they
00:21:00.880Byron Mackay: will give you their insights on what
00:21:02.320Byron Mackay: actually is happening in the world. Um,
00:21:04.559Byron Mackay: and finally, we have hiring partners. We
00:21:07.120Byron Mackay: have hiring partners, which uh at
00:21:09.840Byron Mackay: Gauntlet are the cohort doesn't
00:21:11.280Byron Mackay: interface with the hiring partners as
00:21:12.559Byron Mackay: much, but I still include them because
00:21:15.600Byron Mackay: I'll talk to the hiring partners and
00:21:17.120Byron Mackay: I'll hear what their needs are. I'll
00:21:18.559Byron Mackay: hear what their concerns are, and then
00:21:20.720Byron Mackay: that's something I can take back and
00:21:22.400Byron Mackay: say, "Hey everybody, let's let's work
00:21:24.559Byron Mackay: this out. Let's figure out this the
00:21:26.320Byron Mackay: answer to this problem. How do we
00:21:27.520Byron Mackay: address this in a company?" Um, so we
00:21:30.240Byron Mackay: kind of get this feedback loop from them
00:21:32.480Byron Mackay: that then gets incorporated into our
00:21:34.720Byron Mackay: ecosystem and our learning environment
00:21:37.120Byron Mackay: and our and and our curriculum. Um, so
00:21:40.400Byron Mackay: it just makes for a really cohesive uh
00:21:43.919Byron Mackay: environment where everybody is uh is is
00:21:47.679Byron Mackay: constantly pushing one another and
00:21:49.200Byron Mackay: helping one another and expanding what
00:21:51.440Byron Mackay: we already know um into into other
00:21:53.840Byron Mackay: areas. So the tribe, the group of people
00:21:56.960Byron Mackay: you surround yourself with, the
00:21:58.159Byron Mackay: different resources, the different
00:21:59.679Byron Mackay: personnel that you can interact with and
00:22:02.320Byron Mackay: the value that they bring to your
00:22:03.760Byron Mackay: learning, that can go a long way. This
00:22:06.159Byron Mackay: can be found also in Slack groups. It
00:22:08.480Byron Mackay: can be found in, you know, Reddit
00:22:10.159Byron Mackay: threads, um, meetups in your local area,
00:22:13.919Byron Mackay: um, of course at your own place of
00:22:15.840Byron Mackay: employment, friends, um, you know, going
00:22:19.440Byron Mackay: to networking events even. you can talk
00:22:22.000Byron Mackay: to some people and get some interesting
00:22:23.520Byron Mackay: insights there. Um, so the all different
00:22:27.600Byron Mackay: ways that you can build that tribe
00:22:30.240Byron Mackay: around you and uh and get some great
00:22:32.960Byron Mackay: insights and that's going to help you go
00:22:34.320Byron Mackay: a long way. So remember we talked about
00:22:37.360Byron Mackay: uh doing um we talked about doing with
00:22:40.159Byron Mackay: feedback
00:22:41.840Byron Mackay: um and how this tribe can can help
00:22:44.400Byron Mackay: provide that feedback in numerous ways
00:22:47.039Byron Mackay: to really get you going. And I'll just
00:22:49.360Byron Mackay: highlight one more time the rate of
00:22:51.280Byron Mackay: success in learning with a good tribe.
00:22:54.559Byron Mackay: Um it goes it it's significantly more
00:22:58.080Byron Mackay: successful um than without. And I think
00:23:01.360Byron Mackay: about all the great engineers I know,
00:23:03.280Byron Mackay: they all have this kind of mentality of
00:23:05.600Byron Mackay: building their tribe. It's it's really
00:23:06.960Byron Mackay: quite impressive. Um and that's one
00:23:09.280Byron Mackay: thing at Gauntlet that we feel makes us
00:23:11.520Byron Mackay: very successful is is we provide a lot
00:23:14.400Byron Mackay: of that environment here for you. Um and
00:23:17.039Byron Mackay: and we've seen again in my career I've
00:23:19.200Byron Mackay: done both and uh this is by far the way
00:23:21.760Byron Mackay: to go and we're going to keep it that
00:23:22.960Byron Mackay: way.
00:23:25.520Byron Mackay: Um
00:23:27.360Byron Mackay: so one of the one of the things that uh
00:23:31.440Byron Mackay: we do here at Gauntlet, this is one of
00:23:32.960Byron Mackay: the more recent changes that we have. Um
00:23:35.840Byron Mackay: but we share ideas constantly here
00:23:37.440Byron Mackay: because you're with this group of people
00:23:38.799Byron Mackay: who are all trying to understand this AI
00:23:42.799Byron Mackay: space and it's changing all the time.
00:23:45.120Byron Mackay: Um, we want to share ideas. We want to
00:23:47.679Byron Mackay: share what we're working on, right? We
00:23:49.280Byron Mackay: want to get that feedback. And that's
00:23:51.360Byron Mackay: not new to G. That's always been there
00:23:52.880Byron Mackay: since day one because everyone's always
00:23:54.400Byron Mackay: talking, always chatting about their
00:23:55.679Byron Mackay: stuff. But we've actually kind of
00:23:58.400Byron Mackay: encapsulated that into what we call an
00:24:00.960Byron Mackay: architecture defense. And what an
00:24:03.840Byron Mackay: architecture defense consists of is five
00:24:07.440Byron Mackay: of you, five people will have will be
00:24:10.000Byron Mackay: given an assignment and it's a case
00:24:11.919Byron Mackay: study. And these five people will take
00:24:13.919Byron Mackay: this case study and they'll all come up
00:24:15.600Byron Mackay: with their own solutions to this case
00:24:17.840Byron Mackay: study. And then within hours of
00:24:20.960Byron Mackay: receiving the assignment and creating
00:24:22.640Byron Mackay: that plan, they will sit in a room
00:24:25.039Byron Mackay: together and each one takes a turn
00:24:27.120Byron Mackay: presenting their plan.
00:24:30.480Byron Mackay: And they get five minutes. And then
00:24:33.120Byron Mackay: after they get they give their
00:24:34.400Byron Mackay: five-minute presentation, the rest of
00:24:36.240Byron Mackay: the group gets five minutes to poke
00:24:37.840Byron Mackay: holes, critique, ask all the questions
00:24:39.440Byron Mackay: they want.
00:24:40.960Byron Mackay: It can be grueling. It can be um it can
00:24:45.279Byron Mackay: be difficult to stand up there in front
00:24:47.840Byron Mackay: of peers, but they're giving you very
00:24:50.799Byron Mackay: valuable feedback in those sessions. And
00:24:52.960Byron Mackay: something that I'm very proud of to say
00:24:55.279Byron Mackay: has been um very successful. Sometimes
00:24:57.840Byron Mackay: it doesn't feel that way cuz it's it's
00:24:59.440Byron Mackay: it can feel like you're getting put on
00:25:01.440Byron Mackay: the spot and and rightfully so you are.
00:25:04.159Byron Mackay: Um but the outputs that we get from it,
00:25:08.640Byron Mackay: the the changes that are then made and
00:25:11.200Byron Mackay: the direction that people are going is
00:25:14.000Byron Mackay: uh has has had a remarkable difference
00:25:17.200Byron Mackay: since we started doing those. Um so
00:25:21.200Byron Mackay: that's just one one way that we're kind
00:25:23.279Byron Mackay: of iterating on these ideas of how we
00:25:25.120Byron Mackay: get feedback.
00:25:28.799Byron Mackay: Okay.
00:25:30.480Byron Mackay: Um I want to talk about uh I want to
00:25:34.080Byron Mackay: talk about this here with learning with
00:25:36.080Byron Mackay: LMS because
00:25:39.120Byron Mackay: it wasn't you know this is a new thing
00:25:41.039Byron Mackay: that we have and it's it's really
00:25:42.960Byron Mackay: powerful and I don't know if I'm sure
00:25:44.880Byron Mackay: many of you have if not all of you have
00:25:47.279Byron Mackay: sat down with an element said I don't
00:25:48.640Byron Mackay: know what this is explain to me like I'm
00:25:50.159Byron Mackay: five and suddenly you get this great
00:25:52.080Byron Mackay: explanation and you can go back and ask
00:25:53.760Byron Mackay: questions and be like what about in this
00:25:55.200Byron Mackay: case or in my life I do this and is this
00:25:57.600Byron Mackay: you know is this related Um,
00:26:01.039Byron Mackay: and it's it's it's amazing, right? Like
00:26:03.600Byron Mackay: you're not reading blog posts. You're
00:26:05.039Byron Mackay: not trying to find the thing that um
00:26:07.200Byron Mackay: that works just for you. Uh, and we
00:26:10.799Byron Mackay: encourage it. We encourage learning with
00:26:12.400Byron Mackay: LLMs. I will highlight again though that
00:26:14.799Byron Mackay: it is only a piece of the process and it
00:26:17.520Byron Mackay: has to be done right. It really does. Um
00:26:21.360Byron Mackay: there's there are learning strategies
00:26:23.120Byron Mackay: with LLMs that are being used and
00:26:26.720Byron Mackay: leveraged that uh accelerate the
00:26:30.960Byron Mackay: um the comprehension far better than
00:26:33.760Byron Mackay: we've been able to do it before. U
00:26:35.919Byron Mackay: Miriam, this kind of comes back to to
00:26:37.520Byron Mackay: your question a bit on the consuming
00:26:39.039Byron Mackay: side and how we cannot get too heavily
00:26:41.840Byron Mackay: focused on it, but I'm going to
00:26:43.440Byron Mackay: highlight here that it's still very very
00:26:44.799Byron Mackay: important. Um, I want to talk about a
00:26:47.120Byron Mackay: Harvard study that happened. Um,
00:26:49.679Byron Mackay: physical science fall of 2023. ChachiBT
00:26:53.840Byron Mackay: just came out not too long ago and a
00:26:56.240Byron Mackay: Harvard professor decided, I'm going to
00:26:57.760Byron Mackay: use this, but I want to see how
00:26:59.520Byron Mackay: effective this actually is. So, he built
00:27:01.440Byron Mackay: a chatbot. He built a chatbot and this
00:27:05.039Byron Mackay: chatbot had all of his material and he
00:27:08.799Byron Mackay: made sure he he wanted to make sure that
00:27:10.400Byron Mackay: there were no hallucinations. So, he
00:27:11.600Byron Mackay: gave it a lot of context. These were
00:27:13.600Byron Mackay: older LLMs that were much more highly
00:27:15.919Byron Mackay: likely to hallucinate. Um, and he put
00:27:20.320Byron Mackay: these chatbots he put a chatbot up for
00:27:22.480Byron Mackay: his students to learn from. Now, he had
00:27:24.799Byron Mackay: he always had a control. So, he always
00:27:26.400Byron Mackay: had one class that didn't have the
00:27:27.760Byron Mackay: tutor, the AI tutor, and another class
00:27:29.360Byron Mackay: that did. Um, and consequently, he would
00:27:31.840Byron Mackay: actually flip them. So, one week one
00:27:33.600Byron Mackay: would have him and the next week the
00:27:34.559Byron Mackay: other one would have him back and forth.
00:27:36.799Byron Mackay: And what he noticed was that those the
00:27:39.440Byron Mackay: the weeks that the student had the LLM
00:27:41.840Byron Mackay: tutor, they reported higher engagement
00:27:44.480Byron Mackay: in class. He reported higher grades
00:27:46.799Byron Mackay: overall, better explanations on their
00:27:48.480Byron Mackay: papers. The students that had the AI
00:27:50.960Byron Mackay: were doing far better. Now,
00:27:54.159Byron Mackay: what does that mean for us? Well, it
00:27:56.480Byron Mackay: means that this works. It means there's
00:27:58.000Byron Mackay: a way that learning with LMS is actually
00:28:00.080Byron Mackay: really impactful. It gets us up to
00:28:02.399Byron Mackay: speed. Remember what I said earlier
00:28:03.760Byron Mackay: about LLM will give you the average? So,
00:28:06.399Byron Mackay: we're going to get to the average a lot
00:28:07.919Byron Mackay: faster if we just can leverage them
00:28:10.159Byron Mackay: correctly. And then from there, we can
00:28:12.240Byron Mackay: actually springboard off into different
00:28:13.600Byron Mackay: directions. And of course, LMS can help
00:28:15.840Byron Mackay: us do web searches and research papers
00:28:17.520Byron Mackay: and find other information. And that's
00:28:18.799Byron Mackay: all great and good. But it can speed up
00:28:21.279Byron Mackay: our learning quite a bit or at least our
00:28:22.799Byron Mackay: understanding or our our breath of
00:28:25.760Byron Mackay: awareness is what I how I like to put
00:28:27.760Byron Mackay: it. Um, I have something that I can I'm
00:28:31.120Byron Mackay: happy to share with everybody here. Um,
00:28:33.440Byron Mackay: it's a resource that I shared in my last
00:28:35.120Byron Mackay: night's school where it's a it's a um
00:28:39.679Byron Mackay: it's a skill I've turned into a skill
00:28:42.080Byron Mackay: now where it'll help you learn system
00:28:44.640Byron Mackay: design, but it does it in a way where it
00:28:47.840Byron Mackay: has a structure, but it quizzes you and
00:28:50.559Byron Mackay: it gives you case studies and does
00:28:52.000Byron Mackay: different things to help kind of
00:28:53.760Byron Mackay: scaffold your learning.
00:28:56.080Byron Mackay: um when we use LMS with some scaffold
00:28:58.480Byron Mackay: learning that's something that will
00:29:00.399Byron Mackay: really help us to um accelerate our
00:29:03.679Byron Mackay: understanding and then we can go and
00:29:06.880Byron Mackay: implement it. Um I really like the idea
00:29:09.919Byron Mackay: of a of a college classroom where we go
00:29:12.000Byron Mackay: home and study with an LLM. Um maybe we
00:29:14.799Byron Mackay: even read and and reading books as well,
00:29:17.120Byron Mackay: but we kind of study on our own. We come
00:29:19.120Byron Mackay: to class, we're ready, we have this
00:29:20.880Byron Mackay: great discussion. We start to poke holes
00:29:23.760Byron Mackay: in each other's ideas and see where
00:29:25.120Byron Mackay: things work and things don't. And then
00:29:26.720Byron Mackay: we can go build and then we come back
00:29:28.399Byron Mackay: and get feedback. Um, and then we kind
00:29:30.480Byron Mackay: of iterate there between the build and
00:29:31.919Byron Mackay: the feedback and, you know, incorporate
00:29:33.760Byron Mackay: the element. There you have that whole
00:29:35.840Byron Mackay: system in place and you're going to go a
00:29:38.159Byron Mackay: whole lot further. Now, you do that with
00:29:39.840Byron Mackay: a group of people who are like-minded
00:29:41.200Byron Mackay: and want to do the same thing. You're
00:29:42.880Byron Mackay: going to go even further and you see how
00:29:45.120Byron Mackay: this all kind of builds up together. Um,
00:29:48.320Byron Mackay: if you have the right balance between
00:29:50.640Byron Mackay: those, then you can really accelerate
00:29:52.960Byron Mackay: your learning and really push yourself.
00:29:55.360Byron Mackay: Um, this is actually a product that
00:29:56.799Byron Mackay: we're working on internally for our
00:29:58.640Byron Mackay: students. We have a scaffold AI tutor.
00:30:01.200Byron Mackay: We call it Traverse and it's a tool that
00:30:03.919Byron Mackay: does exactly that. We have things that
00:30:06.080Byron Mackay: we want uh our cohorts to understand. Um
00:30:10.159Byron Mackay: that expands their breath of of
00:30:12.159Byron Mackay: awareness and uh it gives them just
00:30:14.880Byron Mackay: enough to say I can go and articulate
00:30:17.679Byron Mackay: how to solve a problem. Even if I
00:30:20.000Byron Mackay: haven't solved that problem before, I'm
00:30:21.600Byron Mackay: at least aware of the solutions that
00:30:23.120Byron Mackay: exist and I can start to um iterate on
00:30:25.679Byron Mackay: those. So um something anyway it's it's
00:30:31.360Byron Mackay: it's it's something that we're working
00:30:33.440Byron Mackay: on. we're we're iterating on with our
00:30:35.200Byron Mackay: with our students as we speak. So, um
00:30:37.919Byron Mackay: it's a it's a fun fun opportunity to see
00:30:40.480Byron Mackay: just how far these LMS can can take some
00:30:42.640Byron Mackay: learning.
00:30:45.520Byron Mackay: Okay.
00:30:47.440Byron Mackay: Now, what about curriculum? Now, here's
00:30:50.799Byron Mackay: here's all the pieces that shape the
00:30:53.440Byron Mackay: environment, but what about curriculum?
00:30:56.240Byron Mackay: Curriculum was the smallest piece.
00:30:57.679Byron Mackay: Remember the formal training, the 10%.
00:31:01.039Byron Mackay: Is that really needed? Do we actually
00:31:02.399Byron Mackay: need it? We do. We do need it. Why?
00:31:05.679Byron Mackay: Because it gives us focus. It gives us a
00:31:09.279Byron Mackay: path that we can follow. It's there's
00:31:12.399Byron Mackay: not this stop and going, "Okay, where
00:31:13.760Byron Mackay: would we go next?" No, we know where
00:31:14.960Byron Mackay: we're going next. It keeps us moving.
00:31:16.880Byron Mackay: Keeps us moving in the right direction.
00:31:18.640Byron Mackay: And we have these checkpoints where we
00:31:20.159Byron Mackay: can come in and talk about different
00:31:21.520Byron Mackay: topics and then we move on to the next
00:31:23.360Byron Mackay: one. That structured learning can go a
00:31:24.880Byron Mackay: long way. Can really help us um learn
00:31:27.760Byron Mackay: and do and build things.
00:31:30.240Byron Mackay: um it's a it's still a very important
00:31:33.679Byron Mackay: piece, albeit a smaller piece to the to
00:31:36.559Byron Mackay: the whole, but if we don't have that,
00:31:39.840Byron Mackay: it's we we we can lose our direction
00:31:42.960Byron Mackay: real quick. Um and so it's it's still a
00:31:45.760Byron Mackay: critical critical piece to the whole
00:31:47.519Byron Mackay: thing. But if we have curriculum, we
00:31:49.919Byron Mackay: talked about our tribe. Our tribe being
00:31:52.320Byron Mackay: where the people we surround ourselves
00:31:54.399Byron Mackay: with, they can give us feedback, they
00:31:56.320Byron Mackay: can give us new ideas. our our our
00:31:59.760Byron Mackay: problems be become shared problems and
00:32:02.240Byron Mackay: we can solve them um together as a team
00:32:05.519Byron Mackay: because if one person's pro having the
00:32:07.039Byron Mackay: problem, somebody else uh is likely
00:32:09.440Byron Mackay: having the same issue. Um we talked
00:32:13.200Byron Mackay: about building over consuming. We still
00:32:15.200Byron Mackay: need to consume
00:32:17.200Byron Mackay: um but we need to build a lot. Remember
00:32:19.840Byron Mackay: what we talked about on the job training
00:32:21.360Byron Mackay: was 70%. we need to be doing something
00:32:23.919Byron Mackay: with what we are learning or it won't
00:32:26.240Byron Mackay: happen or it just it just won't uh won't
00:32:28.799Byron Mackay: stick as well and we won't grow and
00:32:30.799Byron Mackay: expand and find out what we what we
00:32:33.200Byron Mackay: don't know. Um that's one thing here at
00:32:36.080Byron Mackay: Glet. We always have projects we're
00:32:37.679Byron Mackay: working on. Always, always, always. And
00:32:40.000Byron Mackay: they're hefty. They're big. Um we talk
00:32:43.600Byron Mackay: about how a Gauntlet cohort there's uh
00:32:46.960Byron Mackay: it's the amount of time you spend in a
00:32:49.279Byron Mackay: week is anywhere between 80 to 100
00:32:50.720Byron Mackay: hours. And that's not just because we
00:32:54.080Byron Mackay: say stay at your computer for 80 to 100
00:32:55.840Byron Mackay: hours. We give you really hefty projects
00:32:58.080Byron Mackay: that take a lot of time. Um, and we want
00:33:01.200Byron Mackay: you sp we we we probably spend um we
00:33:04.720Byron Mackay: spend at least 70% of our time building
00:33:08.159Byron Mackay: and working on on our projects. So we
00:33:12.159Byron Mackay: have that building side. Uh then we have
00:33:15.039Byron Mackay: teaching and sharing of ideas that can
00:33:17.200Byron Mackay: be structured time for feedback. It can
00:33:20.080Byron Mackay: just be talking to your different cohort
00:33:21.679Byron Mackay: members. It can be posting something in
00:33:23.360Byron Mackay: Slack and getting some help. Um, there's
00:33:26.559Byron Mackay: a lot of opportunities for teaching and
00:33:29.919Byron Mackay: sharing your own ideas and receiving
00:33:32.000Byron Mackay: those ideas and teachings from others
00:33:34.240Byron Mackay: while you're while you're here. And of
00:33:36.320Byron Mackay: course, you can find that elsewhere as
00:33:38.000Byron Mackay: well, but I think it's a really critical
00:33:39.600Byron Mackay: piece uh in any learning environment to
00:33:42.320Byron Mackay: really accelerate learning. And finally,
00:33:44.480Byron Mackay: LLMs. They have really influenced and
00:33:49.360Byron Mackay: impacted how quickly we can learn
00:33:51.679Byron Mackay: things. And uh I I'll be my personal
00:33:55.600Byron Mackay: opinion, the world hasn't figured it out
00:33:57.840Byron Mackay: yet. I I think they're scared. I think a
00:34:00.960Byron Mackay: lot of people are scared of what these
00:34:02.559Byron Mackay: what LLMs are capable of. And I think a
00:34:05.039Byron Mackay: lot of people are, you know, keeping
00:34:07.519Byron Mackay: keeping their distance, which I think is
00:34:10.560Byron Mackay: I think I'm talking speaking to the
00:34:11.919Byron Mackay: choir here quite a bit. That is the
00:34:14.079Byron Mackay: wrong thing to do.
00:34:15.679Byron Mackay: Don't stand off and and don't let them
00:34:17.839Byron Mackay: don't let them scare you. Like leverage
00:34:19.520Byron Mackay: them. They're great. They're amazing
00:34:20.800Byron Mackay: tools. Um I have learned so much and
00:34:23.760Byron Mackay: I've expanded my my own knowledge so
00:34:26.320Byron Mackay: much because I have um put together a
00:34:30.560Byron Mackay: structured chatbot inside of inside of
00:34:32.639Byron Mackay: Claude that I can use to start learning
00:34:35.040Byron Mackay: and understanding things. Um there's a
00:34:37.679Byron Mackay: lot of opportunity there to uh um to
00:34:41.359Byron Mackay: expand what we know and really really
00:34:43.040Byron Mackay: push ourselves. um beyond what we what
00:34:45.599Byron Mackay: we've been capable of before and so much
00:34:47.520Byron Mackay: faster than before. Um it's really it's
00:34:50.879Byron Mackay: really quite impressive and amazing.
00:34:54.720Byron Mackay: All right.
00:34:57.040Byron Mackay: So in summary here, um the lesson plan
00:35:02.320Byron Mackay: is only 10%.
00:35:05.520Byron Mackay: The environment you create is the other
00:35:08.560Byron Mackay: 90%.
00:35:10.720Byron Mackay: That's how critical the environment is.
00:35:13.839Byron Mackay: Um, if you have the right environment,
00:35:17.040Byron Mackay: if you can create that environment for
00:35:18.720Byron Mackay: yourself or you find a group of
00:35:20.400Byron Mackay: like-minded individuals,
00:35:22.640Byron Mackay: this is going to push you more um, and
00:35:27.040Byron Mackay: get you get you moving faster than ever
00:35:29.280Byron Mackay: before. A lesson plan is important. It's
00:35:31.920Byron Mackay: critical to this. It doesn't it doesn't
00:35:33.599Byron Mackay: work well without one. But if you can
00:35:36.160Byron Mackay: put these two together and you can have
00:35:37.920Byron Mackay: that have that environment, there's no
00:35:41.440Byron Mackay: stopping you. That's one thing about uh
00:35:43.760Byron Mackay: here at Gauntlet, we're not afraid to
00:35:45.119Byron Mackay: tell people what we talk about. I can
00:35:46.560Byron Mackay: give you the syllabus. I can I could sit
00:35:49.040Byron Mackay: down and say, "Here's a list of
00:35:50.800Byron Mackay: everything you should go through, but
00:35:52.880Byron Mackay: it's only 10% of the of of what of of
00:35:56.400Byron Mackay: how you can really push yourself off and
00:35:58.640Byron Mackay: become effective." Now, obviously, I'm
00:36:02.000Byron Mackay: going to encourage this, but I really uh
00:36:06.079Byron Mackay: really want to tell you that it's a it's
00:36:08.320Byron Mackay: my honest assessment
00:36:10.560Byron Mackay: um from the things that I've seen and
00:36:12.560Byron Mackay: the experiences I've had, the people
00:36:14.000Byron Mackay: that have gone through the program that
00:36:16.720Byron Mackay: Gauntlet is a great environment to work
00:36:20.160Byron Mackay: in and and learn in. It's rigorous. It's
00:36:24.240Byron Mackay: got a amazing tribe of people. um you
00:36:27.760Byron Mackay: will consume, you will build far more
00:36:30.640Byron Mackay: than you have. Um it'll be a chance for
00:36:33.920Byron Mackay: you to really expand and uh the breadth
00:36:37.920Byron Mackay: of what you know and how deep you know
00:36:39.839Byron Mackay: things and it's going to push you in so
00:36:42.160Byron Mackay: many ways to do that. And it's not just
00:36:44.320Byron Mackay: what we teach. It's about the people
00:36:47.119Byron Mackay: that you're around and the people that
00:36:49.520Byron Mackay: are here to support you. not just
00:36:51.440Byron Mackay: gauntlet staff but also the other people
00:36:53.440Byron Mackay: that come in that are trying to achieve
00:36:54.800Byron Mackay: the same thing you are. You put all that
00:36:57.839Byron Mackay: together and it's a recipe for success.
00:37:01.359Byron Mackay: And so I would hope that um that anybody
00:37:05.520Byron Mackay: who is interested or or thinking how can
00:37:08.079Byron Mackay: I get myself beyond where I am and and
00:37:11.040Byron Mackay: going off to you know get myself into a
00:37:13.040Byron Mackay: better position for my career or how do
00:37:15.200Byron Mackay: I navigate the AI era that we're in so
00:37:18.880Byron Mackay: that I make sure my my job is secure or
00:37:21.599Byron Mackay: I'm still relevant. This is a great
00:37:24.160Byron Mackay: place to come. We really we have focused
00:37:27.280Byron Mackay: ourselves on creating an amazing
00:37:29.520Byron Mackay: environment that you can learn in. Um
00:37:32.560Byron Mackay: and I hope that uh that if any of you uh
00:37:35.839Byron Mackay: have a will that all of you will reach
00:37:37.920Byron Mackay: out um and and see what it takes to get
00:37:41.040Byron Mackay: in and we can see you in class here at
00:37:43.040Byron Mackay: Gauntlet.
00:37:45.200Byron Mackay: Okay,
00:37:47.599Byron Mackay: that's everything I have today uh for
00:37:50.079Byron Mackay: you. Uh, I want to pause here for some
00:37:52.880Byron Mackay: questions um that I can answer. Um, feel
00:37:56.720Byron Mackay: free to ask anything about Gauntlet,
00:37:59.040Byron Mackay: anything about learning, um, resources.
00:38:01.839Byron Mackay: I'm happy to to give you resources for
00:38:04.560Byron Mackay: your own personal plans, uh, or topics
00:38:07.599Byron Mackay: that you may want or anything about the
00:38:09.200Byron Mackay: market that I can shed some light on
00:38:11.119Byron Mackay: with what I do know about, uh, from from
00:38:14.000Byron Mackay: talking to our hiring partners or what
00:38:16.880Byron Mackay: I've seen our graduates go off and do.
00:38:19.520Byron Mackay: um what I see in the AI uh environment,
00:38:23.760Byron Mackay: anything under there, anything else
00:38:25.520Byron Mackay: beyond, happy to take a whack at it. Um
00:38:29.280Byron Mackay: Unice asked, "What email do we reach out
00:38:31.359Byron Mackay: to for Gauntlet?" Uh I would reach out,
00:38:33.920Byron Mackay: well, there's the Gauntlet website. Um
00:38:36.640Byron Mackay: there's the Gauntlet website that we can
00:38:38.400Byron Mackay: get you uh pointed to and uh that'll be
00:38:41.680Byron Mackay: gauntletai.com. That's a great place to
00:38:43.839Byron Mackay: get to to fill out your application and
00:38:45.839Byron Mackay: get going. Um, and then we will be
00:38:48.880Byron Mackay: reaching out to you and get you uh to
00:38:52.320Byron Mackay: help you get through the application
00:38:53.520Byron Mackay: process and make sure we can get you uh
00:38:55.920Byron Mackay: get you going here. Um, another question
00:38:59.040Byron Mackay: here. I have a quick question on the
00:39:00.400Byron Mackay: application. I applied on the for the
00:39:01.839Byron Mackay: July 6 cohort. My application says still
00:39:04.880Byron Mackay: under review. Okay, that's a that's a
00:39:07.760Byron Mackay: great call. I'm going to take your name
00:39:09.760Byron Mackay: back to our admissions guy and make sure
00:39:12.640Byron Mackay: that uh that he is aware. And so I'm
00:39:15.680Byron Mackay: gonna take a quick note here and just
00:39:19.440Byron Mackay: make sure that he's uh
00:39:22.880Byron Mackay: he is uh make sure he catches your name.
00:39:25.520Byron Mackay: He we have a lot of applicants that come
00:39:27.280Byron Mackay: in and so uh it's it's it's uh it's very
00:39:32.240Byron Mackay: likely I just need a poke prod and poke
00:39:34.000Byron Mackay: and we'll get you going. Um same
00:39:37.119Byron Mackay: question regarding application still
00:39:38.480Byron Mackay: pending. All right, I'll take your name
00:39:39.680Byron Mackay: down as well. Get yours going there.
00:39:44.079Byron Mackay: Um Nathan asked what interested in what
00:39:46.720Byron Mackay: gauntlets our graduates end up doing
00:39:48.240Byron Mackay: after the cohort. So part of gauntlet I
00:39:50.640Byron Mackay: didn't mention this earlier but part of
00:39:52.320Byron Mackay: gauntlet is that we bring in hiring
00:39:54.960Byron Mackay: partners who are looking for people um
00:39:57.760Byron Mackay: with the skill sets that we teach and
00:40:00.160Byron Mackay: who have proven themselves that they can
00:40:02.320Byron Mackay: learn in a rigorous environment and stay
00:40:03.920Byron Mackay: up to date with their AI skills. And so
00:40:07.359Byron Mackay: uh one what uh uh so what our graduates
00:40:11.760Byron Mackay: I'm doing is getting jobs with these
00:40:12.960Byron Mackay: with these hiring partners. We actually
00:40:14.320Byron Mackay: bring them in and they get interviewed
00:40:16.480Byron Mackay: in person. Um part of Gauntlet is that
00:40:19.280Byron Mackay: we fly you out to Austin, Austin, Texas,
00:40:21.280Byron Mackay: and we're all in the office together. We
00:40:22.800Byron Mackay: have this great office with um all the
00:40:25.359Byron Mackay: food and snacks and um and desk space
00:40:28.800Byron Mackay: and monitors um that you can you can ask
00:40:31.520Byron Mackay: for. We do your laundry, pay for all
00:40:33.359Byron Mackay: your meals. It's great. And then we
00:40:34.800Byron Mackay: bring in hiring partners come interview
00:40:36.079Byron Mackay: you. So they all end up getting jobs and
00:40:38.160Byron Mackay: moving on from there. Um so it's really
00:40:40.480Byron Mackay: a a fun a fun time. Really exciting to
00:40:43.520Byron Mackay: see when everybody graduates and where
00:40:44.800Byron Mackay: they go off to. Um
00:40:48.079Byron Mackay: one asked what's the passing score for
00:40:49.599Byron Mackay: the CCAT? Yes, the CCAT is our uh is a
00:40:52.720Byron Mackay: cognitive test that we have everybody
00:40:54.240Byron Mackay: take. 40. 40 is the number um that we're
00:40:57.839Byron Mackay: going for with the with the CCAT. Um,
00:41:00.960Byron Mackay: it's uh if you haven't taken the CCAP
00:41:02.800Byron Mackay: before, it's I wouldn't say the
00:41:04.319Byron Mackay: questions are necessarily difficult, but
00:41:06.640Byron Mackay: it is a time test and they don't give
00:41:08.160Byron Mackay: you a lot of time. So, it's more about
00:41:09.440Byron Mackay: speed and doing your best in that time
00:41:12.079Byron Mackay: frame. Um, so just a little hint there
00:41:14.560Byron Mackay: that you get about 16, 17, 18 seconds
00:41:17.680Byron Mackay: per question. Um, that's about as much
00:41:20.240Byron Mackay: time as you want to give yourself on
00:41:21.839Byron Mackay: that test. It's a it's a tough one. Uh,
00:41:24.480Byron Mackay: just in terms of managing time, not
00:41:26.079Byron Mackay: necessarily in terms of questions.
00:41:28.480Byron Mackay: Um, all right. Nathan asks specifically,
00:41:31.680Byron Mackay: do they typically take uh AI engineering
00:41:34.319Byron Mackay: roles or more traditional software
00:41:35.760Byron Mackay: engineering roles? They're they're more
00:41:37.839Byron Mackay: AI engineering based. Um,
00:41:41.280Byron Mackay: they are there there's there's this kind
00:41:43.119Byron Mackay: of mix you could say when these roles
00:41:46.319Byron Mackay: where it's um part of it is is of course
00:41:49.440Byron Mackay: traditional software engineering, but
00:41:51.440Byron Mackay: they always want to have that AI
00:41:53.040Byron Mackay: component as well. Whether that's your
00:41:55.520Byron Mackay: coding workflow is a as a Gentic or your
00:41:58.640Byron Mackay: um you can build AI products um or you
00:42:01.200Byron Mackay: can be a leader at their company for um
00:42:04.560Byron Mackay: for implementing AI. It's it's those
00:42:07.119Byron Mackay: kinds of things that they're that
00:42:08.240Byron Mackay: they're looking for.
00:42:10.400Byron Mackay: All right. Ash asked uh also same thing
00:42:13.599Byron Mackay: for me. Okay. Yep. I'll put your name in
00:42:16.560Byron Mackay: um and we'll get you uh get that over to
00:42:19.040Byron Mackay: our uh to our applications.
00:42:22.800Byron Mackay: Um,
00:42:24.720Byron Mackay: perfect. Um, Unice asked, "Can somebody
00:42:27.359Byron Mackay: apply outside the can someone outside
00:42:29.520Byron Mackay: the US apply?" Um, right now we
00:42:32.560Byron Mackay: everybody needs to be a US citizen. Um,
00:42:35.680Byron Mackay: they need to be able to work in the US
00:42:37.200Byron Mackay: without um a visa. And so, uh, so for if
00:42:41.359Byron Mackay: you're not a US citizen, then Gauntlet
00:42:44.160Byron Mackay: might not be the best fit for you.
00:42:45.680Byron Mackay: Whereas, if you are a US citizen, you're
00:42:47.119Byron Mackay: just living outside the US, that's
00:42:48.480Byron Mackay: perfectly fine as long as you're okay
00:42:50.400Byron Mackay: coming back to the US. Um,
00:42:54.240Byron Mackay: all right. Uh, Stephen asked, "I would
00:42:56.560Byron Mackay: like to join the next Gauntlet cohort. I
00:42:57.920Byron Mackay: applied in the past but didn't pursue
00:42:59.200Byron Mackay: strongly enough. I spent the past few
00:43:01.200Byron Mackay: months building new software and
00:43:02.240Byron Mackay: programs and really enjoy it." Awesome.
00:43:03.920Byron Mackay: Stephen, you're the kind of guy I like
00:43:05.520Byron Mackay: to talk to. Um, I would fill out the
00:43:08.160Byron Mackay: application, start the application
00:43:09.359Byron Mackay: again, uh, or if it lets you continue
00:43:12.000Byron Mackay: it. Not sure you might have to apply for
00:43:13.359Byron Mackay: a new one if it's been long enough, but
00:43:15.440Byron Mackay: uh, get to the website, see what you can
00:43:17.760Byron Mackay: do, and we'll get you going.
00:43:21.280Byron Mackay: Um,
00:43:22.880Byron Mackay: all right. Uh,
00:43:25.040Byron Mackay: Alexandra asks, "I'm early in my career
00:43:27.280Byron Mackay: and still in school, but I'm really
00:43:28.800Byron Mackay: interested in attending. What skills,
00:43:30.640Byron Mackay: experiences, and projects should I focus
00:43:32.640Byron Mackay: on to maximize my chances of being a
00:43:34.640Byron Mackay: strong candidate?" Uh, Alexander, that's
00:43:37.200Byron Mackay: a great question. I would focus on
00:43:41.280Byron Mackay: um building building something really
00:43:44.640Byron Mackay: challenging. I think because you're
00:43:46.640Byron Mackay: doing already in school, I know you're
00:43:48.240Byron Mackay: going to get the all the all the other
00:43:49.680Byron Mackay: things figured out, but what really sets
00:43:51.599Byron Mackay: people apart is
00:43:54.560Byron Mackay: um is digging into a hard problem and
00:43:58.160Byron Mackay: and then figuring it out. Um it doesn't
00:44:00.960Byron Mackay: have to be something so large or so big,
00:44:03.920Byron Mackay: but there are enough hard problems out
00:44:05.440Byron Mackay: there that you can find where it's
00:44:07.920Byron Mackay: building it and building it well and
00:44:09.440Byron Mackay: building it right. You're going to find
00:44:10.880Byron Mackay: hurdles that you have to overcome. And
00:44:13.119Byron Mackay: just that process of overcoming those
00:44:14.800Byron Mackay: hurdles and learning um what hurdles
00:44:18.000Byron Mackay: exist that goes a long long way for
00:44:21.200Byron Mackay: people especially in your in in your uh
00:44:23.760Byron Mackay: scenario at school where you get that
00:44:25.680Byron Mackay: exposure, get that experience and you
00:44:27.680Byron Mackay: can push yourself on. Where do you get
00:44:29.440Byron Mackay: the where do you get that uh where are
00:44:32.319Byron Mackay: you going to go and find those kinds of
00:44:33.680Byron Mackay: problems? Go find some people to talk to
00:44:35.920Byron Mackay: who are a little bit further along in
00:44:37.359Byron Mackay: their careers. Go talk to them. Go see
00:44:39.359Byron Mackay: what they're working on. um show
00:44:41.839Byron Mackay: interest in what they're working on and
00:44:44.160Byron Mackay: build those relationships with them.
00:44:46.319Byron Mackay: Help them become part of your tribe
00:44:48.160Byron Mackay: where they can become the people that
00:44:49.520Byron Mackay: you can go to to chat with uh to throw
00:44:52.800Byron Mackay: ideas uh to to toss your ideas over the
00:44:56.319Byron Mackay: wall and see what they think to get that
00:44:58.480Byron Mackay: feedback. Um big part, right, are all
00:45:01.119Byron Mackay: that learning uh learning environment
00:45:03.440Byron Mackay: that we talked about. finding those
00:45:05.440Byron Mackay: people that would really help you push
00:45:07.119Byron Mackay: along um your abilities and get you
00:45:09.440Byron Mackay: ready um to join and and when you're
00:45:11.599Byron Mackay: ready, we'd love to have you.
00:45:15.040Byron Mackay: Um Stephen said, "I don't think the
00:45:17.040Byron Mackay: website will let me reapply. I'd like
00:45:18.480Byron Mackay: you to recheck out." Stephen, I'll take
00:45:20.160Byron Mackay: your name down, get you in with our
00:45:22.720Byron Mackay: admissions guy. Um and we'll see what we
00:45:26.240Byron Mackay: can do.
00:45:28.480Byron Mackay: All right.
00:45:30.000Byron Mackay: Um, Ash, is there any repos, frameworks,
00:45:32.240Byron Mackay: tooling that you would recommend in
00:45:33.599Byron Mackay: general for agentic workflows uh that uh
00:45:37.280Byron Mackay: that we've all used or lean on at
00:45:39.200Byron Mackay: Gauntlet through the through the
00:45:40.400Byron Mackay: syllabus or have it have you uh just
00:45:42.400Byron Mackay: found through the cohorts kind of rings
00:45:44.480Byron Mackay: true? That's a really interesting
00:45:46.720Byron Mackay: question. Um
00:45:49.440Byron Mackay: because it it evolves so much. But
00:45:53.119Byron Mackay: what's so interesting is it actually
00:45:54.880Byron Mackay: always comes back to like the basic
00:45:57.920Byron Mackay: principles that we picked up on early
00:46:00.640Byron Mackay: early on of of how to build agents.
00:46:04.240Byron Mackay: Um spectrum development is effective.
00:46:08.400Byron Mackay: Um,
00:46:11.040Byron Mackay: knowing the tools, knowing the breadth
00:46:12.800Byron Mackay: of tools is really important that you
00:46:14.640Byron Mackay: have available to you. Um, also really
00:46:17.440Byron Mackay: important.
00:46:19.040Byron Mackay: Uh,
00:46:20.720Byron Mackay: but as far as specific frameworks or
00:46:22.800Byron Mackay: tooling, let me put it this way. Um,
00:46:30.160Byron Mackay: the goal of engineering
00:46:33.040Byron Mackay: is to become a manager of agents.
00:46:37.599Byron Mackay: And so
00:46:39.680Byron Mackay: the um your ability to
00:46:44.880Byron Mackay: take a codebase and
00:46:48.160Byron Mackay: instruct it or or add enough context
00:46:51.119Byron Mackay: throughout your codebase that an agent
00:46:52.800Byron Mackay: can go through and follow along is
00:46:55.280Byron Mackay: really critical. Despite the framework
00:46:57.839Byron Mackay: that you use,
00:46:59.839Byron Mackay: the context is context is still king.
00:47:02.560Byron Mackay: Always probably will be. And so when
00:47:06.240Byron Mackay: you're able to um take cla files for
00:47:10.000Byron Mackay: example, I'm a claude person. So cloudmd
00:47:12.480Byron Mackay: files I agent MD files are the ones I
00:47:15.440Byron Mackay: believe for um for other frameworks. But
00:47:18.800Byron Mackay: cloudmd files are basically context you
00:47:20.720Byron Mackay: can drop in different folders in your um
00:47:24.160Byron Mackay: in your codebase. That is the crux of
00:47:28.319Byron Mackay: getting your agent to perform well in
00:47:30.000Byron Mackay: your codebase. You can give it the right
00:47:32.560Byron Mackay: instructions. you can split it up
00:47:34.640Byron Mackay: amongst many cloudmd files across many
00:47:36.960Byron Mackay: different directories so that it's not
00:47:38.400Byron Mackay: just all in one big context dump that
00:47:42.400Byron Mackay: it's very specific and it allows the LM
00:47:44.720Byron Mackay: to focus more on that context and when
00:47:47.920Byron Mackay: you get to a point where you feel like
00:47:49.599Byron Mackay: you've got that down where you've got
00:47:51.920Byron Mackay: that context down where you want it then
00:47:54.880Byron Mackay: you have an agent that just builds that
00:47:57.760Byron Mackay: piece of your application. So, maybe you
00:47:59.839Byron Mackay: have a a backend or database agent or um
00:48:03.839Byron Mackay: a front-end agent and you say, "Okay,
00:48:05.920Byron Mackay: now you're going to go and just build
00:48:07.200Byron Mackay: whatever I tell you." And then you can
00:48:09.040Byron Mackay: tweak that agent a little bit to uh to
00:48:12.560Byron Mackay: your liking until that agent can really
00:48:14.400Byron Mackay: run it autonomously, you know, most of
00:48:16.560Byron Mackay: generally all the time. And then you do
00:48:19.280Byron Mackay: that across the board. You build
00:48:20.400Byron Mackay: different agents for doing different
00:48:21.839Byron Mackay: tasks.
00:48:23.680Byron Mackay: And once you get to the point where you
00:48:25.520Byron Mackay: have all these agents that can run
00:48:27.040Byron Mackay: asynchronously and run autonomously
00:48:30.160Byron Mackay: um and you're not managing them anymore,
00:48:32.240Byron Mackay: you're just kind of giving them tasks.
00:48:33.760Byron Mackay: There's too much task giving for you to
00:48:36.000Byron Mackay: do. That's when you build your
00:48:38.000Byron Mackay: orchestrator above all of them
00:48:40.960Byron Mackay: and then you just talk to the
00:48:42.400Byron Mackay: orchestrator and he distributes the
00:48:44.480Byron Mackay: tasks.
00:48:46.000Byron Mackay: That's where it's all going. That's
00:48:47.599Byron Mackay: where I see everything heading. however
00:48:50.240Byron Mackay: way you look at it, whether it's through
00:48:52.640Byron Mackay: um you know, Ralph loop or whether it's
00:48:55.760Byron Mackay: through um um there's there's there's
00:48:59.440Byron Mackay: other there's other loops that I' that
00:49:00.960Byron Mackay: I've seen or spectrum developments or
00:49:03.119Byron Mackay: all these other things. It all comes
00:49:04.960Byron Mackay: back to can you provide the best context
00:49:07.359Byron Mackay: possible. So, I'm just going to stick
00:49:09.599Byron Mackay: with that principle for now and not give
00:49:11.200Byron Mackay: like a specific framework tooling or
00:49:12.960Byron Mackay: repo. Um
00:49:15.599Byron Mackay: okay, I will give you one repo.
00:49:18.960Byron Mackay: Awesome Skills. Um, that one is great.
00:49:21.839Byron Mackay: Uh, Awesome Skills has a bunch of skills
00:49:23.920Byron Mackay: that you can go through, read through.
00:49:25.359Byron Mackay: It gives you great, um, exposure to what
00:49:28.240Byron Mackay: a good skill looks like and also
00:49:29.599Byron Mackay: probably what some bad skills look like.
00:49:31.359Byron Mackay: U,, but you can leverage those. Um, a
00:49:34.079Byron Mackay: favorite one that keeps coming up is
00:49:35.680Byron Mackay: Grill Me, the Grill Me skill. Um, that
00:49:39.440Byron Mackay: one's excellent because it will really
00:49:41.440Byron Mackay: push you to think through your ideas,
00:49:43.760Byron Mackay: especially if you're doing spectrum
00:49:44.960Byron Mackay: development or you're building a you're
00:49:46.720Byron Mackay: trying to, you know, build an app or or
00:49:48.640Byron Mackay: you have a PR you want to do. The grill
00:49:50.480Byron Mackay: me PR is or the grill me skill is really
00:49:53.200Byron Mackay: good at pointing out things that you
00:49:54.480Byron Mackay: maybe didn't notice before. All right,
00:49:56.720Byron Mackay: long answer. Uh, but there you go. All
00:49:59.760Byron Mackay: right. Um, let's see. Coleman, is this
00:50:03.119Byron Mackay: boot camp a good fit for someone with 10
00:50:04.640Byron Mackay: years of system admin roles but no
00:50:07.359Byron Mackay: traditional dev background building
00:50:09.200Byron Mackay: custom applications with AI um for work
00:50:11.920Byron Mackay: and personally? Um that is a great
00:50:14.880Byron Mackay: question. It depends on what these last
00:50:16.640Byron Mackay: few years you've been able to build I
00:50:18.319Byron Mackay: would say. Um CIS admin roles do not
00:50:22.319Byron Mackay: usually fit well with the um with most
00:50:26.000Byron Mackay: of our hiring partners. Um, but
00:50:28.559Byron Mackay: depending on what you've built, it might
00:50:29.920Byron Mackay: expand to other uh hire partners as
00:50:32.720Byron Mackay: well. So, um, something to look into.
00:50:35.280Byron Mackay: Happy to look at your LinkedIn and have
00:50:37.760Byron Mackay: a conversation or you should just apply.
00:50:40.079Byron Mackay: Um, our our applications guy is great at
00:50:42.800Byron Mackay: helping you kind of navigate that and
00:50:44.240Byron Mackay: make that decision as well.
00:50:46.880Byron Mackay: Um, Derek, how does gauntlet screening
00:50:49.119Byron Mackay: process compare to say IBM's? Is that a
00:50:51.040Byron Mackay: comparison the team has ever made? I
00:50:52.640Byron Mackay: have never made that comparison. That's
00:50:54.800Byron Mackay: a great idea. Um, I'd have to look into
00:50:57.200Byron Mackay: it and see. Um, but very interesting. I
00:51:01.359Byron Mackay: will say that at Gauntlet, we do kind of
00:51:03.680Byron Mackay: see ourselves as like a 10-week
00:51:04.960Byron Mackay: interview. And so, it's a lot of rigor,
00:51:08.319Byron Mackay: um, a lot of learning, a lot of
00:51:09.920Byron Mackay: building, um, a lot of, um, soft skills,
00:51:13.680Byron Mackay: checking, how well do you communicate,
00:51:16.240Byron Mackay: um, how well do you, uh, present ideas,
00:51:18.880Byron Mackay: how well do you give feedback, that sort
00:51:20.559Byron Mackay: of thing. Um, a lot of, uh, a lot of
00:51:24.160Byron Mackay: things that we do there. And then by the
00:51:25.680Byron Mackay: end of that you get hired on by these
00:51:27.520Byron Mackay: companies. So um but like I said you
00:51:31.119Byron Mackay: also get your you get room and board and
00:51:34.079Byron Mackay: meals and a great network and even your
00:51:36.720Byron Mackay: laundry paid for uh while you do it all.
00:51:39.200Byron Mackay: So a lot of fun.
00:51:41.520Byron Mackay: Um
00:51:43.839Byron Mackay: Stephen asks uh question regarding the
00:51:46.000Byron Mackay: projects that uh sponsors or companies
00:51:48.640Byron Mackay: bring to you. Do you have flexibility
00:51:50.079Byron Mackay: for which projects we would like to work
00:51:51.920Byron Mackay: on? I'm particularly interested in
00:51:53.280Byron Mackay: healthcare or chemical engineering. Love
00:51:55.440Byron Mackay: it. Yes, absolutely. We have um we have
00:51:58.559Byron Mackay: partners in both of those and so you'll
00:52:00.800Byron Mackay: have opportunities to work or you could
00:52:02.640Byron Mackay: you would have an opportunity to
00:52:03.599Byron Mackay: interface with them and do projects for
00:52:05.440Byron Mackay: them and talk to them. Um there's a we
00:52:08.160Byron Mackay: have a quite a quite a good range I
00:52:10.559Byron Mackay: would say um of industries that we
00:52:13.040Byron Mackay: cover. Um healthcare and chemical
00:52:14.720Byron Mackay: engineering being two of them.
00:52:17.200Byron Mackay: Um
00:52:18.800Byron Mackay: ah uh Uriel I see you sorry you missed
00:52:22.079Byron Mackay: the session but this is all recorded uh
00:52:24.400Byron Mackay: so you have a chance to go back u which
00:52:26.880Byron Mackay: is your question I just realized is your
00:52:28.800Byron Mackay: question well is it recorded yes it is
00:52:31.359Byron Mackay: um and so you'll be able to get that and
00:52:33.760Byron Mackay: watch it and I believe there will be a
00:52:35.760Byron Mackay: transcript if there's not at least you
00:52:36.960Byron Mackay: get the recording and you'll be able to
00:52:38.240Byron Mackay: watch it there. Um let's see. Um
00:52:44.240Byron Mackay: Ash kind of coming back to my comment
00:52:46.559Byron Mackay: earlier about agentic modularization.
00:52:49.839Byron Mackay: Yes. uh that that is the way it's going
00:52:55.839Byron Mackay: I believe and and I what I see companies
00:52:58.800Byron Mackay: doing
00:53:00.400Byron Mackay: what I see some of the better companies
00:53:02.240Byron Mackay: doing I should say is I see them um I
00:53:06.640Byron Mackay: seeing engineers who have figure who are
00:53:09.359Byron Mackay: figuring that out in the engineering
00:53:11.200Byron Mackay: space and then the engineering team
00:53:13.520Byron Mackay: that's in charge of that the the number
00:53:15.839Byron Mackay: of people gets smaller
00:53:18.720Byron Mackay: because the people they they take some
00:53:21.680Byron Mackay: of those people and start putting them
00:53:22.800Byron Mackay: elsewhere and saying do that again over
00:53:25.359Byron Mackay: here with these people
00:53:28.400Byron Mackay: um because you don't need as many
00:53:30.319Byron Mackay: engineers to maintain it once you get it
00:53:32.160Byron Mackay: set up. But I will say fair warning
00:53:35.280Byron Mackay: getting it set up is not as easy as it
00:53:37.359Byron Mackay: sounds.
00:53:38.880Byron Mackay: It takes a lot of work, a lot of review,
00:53:42.160Byron Mackay: a lot of knowing what good looks like
00:53:44.720Byron Mackay: and
00:53:46.319Byron Mackay: knowing what quality is. Knowing what
00:53:48.640Byron Mackay: quality looks like is going to um
00:53:52.880Byron Mackay: determine how effective the full
00:53:56.480Byron Mackay: orchestration actually is. So, it's a
00:53:59.920Byron Mackay: lot of work, takes takes takes time. Be
00:54:02.240Byron Mackay: patient with it. Um but it will it can
00:54:05.359Byron Mackay: come to fruition. I've I've seen it
00:54:07.680Byron Mackay: myself from different um from different
00:54:10.400Byron Mackay: individuals and within our our group. Um
00:54:13.520Byron Mackay: and it's it's it is possible. It is very
00:54:15.839Byron Mackay: successful and it works. Um, just know
00:54:18.800Byron Mackay: that you are the quality bar. You're the
00:54:20.960Byron Mackay: one measuring the output. So,
00:54:24.960Byron Mackay: awesome. Okay,
00:54:27.520Byron Mackay: that looks like all the questions. Thank
00:54:29.680Byron Mackay: you everybody for coming tonight. Um,
00:54:32.319Byron Mackay: and thank you anybody who's watching
00:54:33.920Byron Mackay: this after the fact. really appreciate
00:54:36.960Byron Mackay: uh you coming and uh I hope that this
00:54:40.079Byron Mackay: has been useful to kind of take into uh
00:54:43.920Byron Mackay: taking in some perspective of what is
00:54:47.040Byron Mackay: required in the learning process.
00:54:48.960Byron Mackay: Remember it's the environment, it's the
00:54:50.800Byron Mackay: people you surround yourself by. It's
00:54:52.880Byron Mackay: how you leverage these tools. Um it's
00:54:55.680Byron Mackay: something that we've found works really,
00:54:57.520Byron Mackay: really well at Gauntlet and I've seen it
00:54:59.760Byron Mackay: work really well throughout my career.
00:55:02.000Byron Mackay: So go out there um build your tribe and
00:55:06.640Byron Mackay: uh and please consider coming to
00:55:09.200Byron Mackay: Gauntlet where we have a great group of
00:55:10.800Byron Mackay: individuals who want to learn just like
00:55:12.400Byron Mackay: you do and we can help jumpst start your
00:55:14.480Byron Mackay: career um with our 10-week program. It's
00:55:17.760Byron Mackay: a it's a lot of fun and I hope to hear
00:55:19.280Byron Mackay: from you all soon. Have a great night
00:55:21.200Byron Mackay: everybody.
FAQ
What is "Why the Environment Beats the Lesson Plan" about? +
What is the 70-20-10 model of learning? +
Why is doing better than just consuming or reading? +
How does feedback accelerate learning? +
Why do cohort-based programs have higher completion rates? +
How do you learn effectively with LLMs? +
How do you get an AI agent to perform well in your codebase? +
CLAUDE.md (or AGENTS.md) files across directories, build specialized agents for specific areas, and add an orchestrator on top once you have more task-giving than you can handle manually.What is an architecture defense at Gauntlet? +
What's next?
Keep building with the rest of Night School, or apply to Gauntlet — ten weeks of technical intensity with the best AI engineers we can find.