文字起こし 🔗
WEBVTT
00:00:00.400 --> 00:00:04.000
Hello and welcome to AI Native DevCon by Tesla.
00:00:04.240 --> 00:00:09.439
This is the fourth edition of this conference, the second in person, and the first in London.
00:00:09.599 --> 00:00:12.960
We're at the brewery in central London near the Barbican Centre.
00:00:13.039 --> 00:00:18.879
And let's go in and take a look at some of the Expo Hall, and we'll give you a flavour of what AI Native DevCon's like.
00:00:19.120 --> 00:00:20.000
Follow me.
00:00:21.280 --> 00:00:27.920
The AI Native Dev is a podcast for developers and engineering leads at the cutting edge of AI and agentic coding.
00:00:28.160 --> 00:00:38.320
Join your hosts, Guy Pajani, and me, Simon Maple, every week as we chat with the most exciting voices in AI and tackle the biggest questions facing developers today.
00:00:38.799 --> 00:00:41.200
This is the AI Native Dev.
00:00:43.679 --> 00:00:47.520
Okay, so here we are at the Expo Centre of AI Native DevCon.
00:00:47.759 --> 00:00:49.600
We're gonna walk in, it's a break, it's busy.
00:00:49.679 --> 00:00:51.359
Let's try and talk to some attendees.
00:00:51.439 --> 00:00:52.399
So follow me.
00:00:52.560 --> 00:00:57.600
And you can see the expo hall is packed, full of uh full of people, full of attendees, and full of sponsors.
00:00:57.759 --> 00:01:00.880
There's areas to get some free shirts and the sponsors going right round.
00:01:01.039 --> 00:01:02.880
So let's find someone to talk to, follow me.
00:01:03.039 --> 00:01:04.079
Tell us a little bit about yourself.
00:01:04.159 --> 00:01:05.519
Uh what's your name and uh who do you work for?
00:01:05.680 --> 00:01:07.200
Uh Sergio Boschek Martin.
00:01:07.439 --> 00:01:11.599
I have a couple of companies, but the reason I'm going here, I have a fintech startup on building alone.
00:01:11.840 --> 00:01:12.319
Oh nice.
00:01:12.480 --> 00:01:15.280
Uh combining AI and blockchain in public and private markets.
00:01:15.599 --> 00:01:25.359
So I'm here to learn uh what are the tools available to make my life easier, to expand the processing, uh to make the production process uh uh output more reliable.
00:01:25.599 --> 00:01:27.599
So look at the tools and practices and processes.
00:01:28.480 --> 00:01:30.480
What are some of the sessions that you've really enjoyed so far?
00:01:30.799 --> 00:01:35.200
Um um um uh Max, the the guy from um Max Fertman?
00:01:35.439 --> 00:01:38.159
Uh no, no, the Max uh Maximilian from from Argentina.
00:01:38.239 --> 00:01:40.959
Uh yeah from Web MCP.
00:01:41.200 --> 00:01:42.079
Yeah, Max Kertman, yeah.
00:01:42.480 --> 00:01:44.480
Uh Max I just remember his last name.
00:01:44.640 --> 00:01:46.000
Uh he was very interesting.
00:01:46.159 --> 00:01:48.079
Then uh Dave Dave Farley.
00:01:48.239 --> 00:01:51.120
Yeah Dave Farley um with NWAV.
00:01:51.280 --> 00:01:57.359
So I'm I'm planning to try all those uh tools uh in my stack and see uh what it can.
00:01:57.920 --> 00:01:58.640
Yeah, very interesting.
00:01:58.799 --> 00:02:16.080
And one of the things I learned with Dave Farley is everyone has a different approach to doing to doing agentic development and agentic coding, and so Dave's some of the things that Dave was saying works obviously really well for him, and then some other people kind of like saying, Archie, I agree with this bit, but not that, so I think it's really important for us to find our own way, but learn from others who are doing this at scale.
00:02:16.800 --> 00:02:26.159
One more session uh that caught my attention yesterday was Matthias Luber from Pi, from the company's different uh I forgot their name, but he's using Pi.
00:02:26.879 --> 00:02:28.479
So I want to try that one as well.
00:02:28.560 --> 00:02:29.520
So amazing, amazing.
00:02:29.599 --> 00:02:30.479
Well enjoy the rest of the conference.
00:02:30.960 --> 00:02:31.439
Thanks for speaking.
00:02:31.759 --> 00:02:32.719
Awesome, thank you.
00:02:36.400 --> 00:02:38.400
So Chris, this is Chris Beatty, one of our speakers.
00:02:38.479 --> 00:02:40.800
Chris, tell us a little bit about yourself and what you presented at uh.
00:02:41.360 --> 00:02:48.719
Yeah, so I've been an engineer for twenty 20 years or so, and what I've done over the last couple of years is I've been building a product team within a consultancy.
00:02:49.039 --> 00:03:00.159
So we've like hired new people, we've tried to build products, and so my talk is about like our lessons learned of adopting AI in that team and like what we're able to achieve, which we wouldn't have been able to achieve, I think, if we tried to do this five five years ago.
00:03:00.400 --> 00:03:00.639
Yeah, yeah.
00:03:00.800 --> 00:03:07.199
And what are the kind of like some of the key learnings that you would say uh if you were to give advice to someone who is on this journey as well?
00:03:07.360 --> 00:03:09.120
What are the key things that you would think could have told you?
00:03:09.680 --> 00:03:16.159
Avoid vanity metrics at all costs, like code commits or how many tokens people are using.
00:03:16.400 --> 00:03:18.560
Have a look at your flow of delivery, right?
00:03:18.719 --> 00:03:21.520
From idea to someone using your piece of software.
00:03:21.759 --> 00:03:25.280
Don't stop at like commit or you know, um outputs.
00:03:25.680 --> 00:03:29.120
Think about the actual outcomes you want to achieve and find the friction, right?
00:03:29.280 --> 00:03:34.719
Because it's not normally in code production, but for our experience, it's not even idea production, like product management.
00:03:34.879 --> 00:03:36.319
We've been pretty good at that now.
00:03:36.479 --> 00:03:42.879
But the the bottleneck for us is actually getting real people to use our software and quickly feeding back on those.
00:03:43.039 --> 00:03:44.000
So you know, yeah.
00:03:44.080 --> 00:03:51.599
And how well are you finding uh, I guess, execs or senior folks within the engineering teams to accept those types of metrics
Load full transcript
over and above token you see?
00:03:51.919 --> 00:03:59.039
Well, actually, like when it comes to it, the higher up you go, if you can talk in the business outcomes, like they love that, right?
00:03:59.439 --> 00:04:00.319
Outcomes over output.
00:04:00.560 --> 00:04:01.039
Yeah, exactly.
00:04:01.919 --> 00:04:02.879
Outcomes over output.
00:04:03.039 --> 00:04:08.400
So I'd say it's normally like I'd say like middle management that are going for vanity metrics because they're easy to measure.
00:04:08.639 --> 00:04:14.319
But if you say, hey, to your execs, to your higher-ups, like we can get these customers doing this by this date.
00:04:14.400 --> 00:04:14.479
Yeah.
00:04:14.639 --> 00:04:16.079
And that's the only measure of success.
00:04:16.639 --> 00:04:26.720
And I don't really care how many tokens you use, like how many commits you make, how many other features that you build, because I think we're all like suffering from building features that are not being used because it's so easy to build features now.
00:04:26.959 --> 00:04:31.199
In fact, if you can improve your outcomes as high as you can with as fewer tokens as you can use, even better.
00:04:31.519 --> 00:04:33.120
And as fewer features in your product, really.
00:04:33.439 --> 00:04:36.560
Because like we've all used I'm on camera, I'm not gonna name a product.
00:04:37.680 --> 00:04:43.360
Um any sessions that you've been to that you've thought have been really amazing, or have you mostly been focusing on the uh on the hallway track?
00:04:43.519 --> 00:04:44.160
Both really.
00:04:44.240 --> 00:04:55.839
I mean, we just went to the session about like measuring the impact of agents and AI development on um organizations and like how people that can become low performers or high performers.
00:04:56.000 --> 00:05:06.800
But overall, I think it's been like a really good conference with a nice balance of like we need a bit of hype, we need to get excited about our work, but there's been a lot of like practical takeaways as well, so people like dealing with reality.
00:05:06.959 --> 00:05:07.600
Yeah, amazing.
00:05:07.680 --> 00:05:11.759
Well, thanks for joining us, thanks for speaking to us and thanks for the supporting conference and talking at uh native dev.
00:05:15.279 --> 00:05:18.480
Tell us a little bit about uh a little bit about yourself, about what you do.
00:05:18.720 --> 00:05:22.720
Uh okay, I work at Chainguard on the uh AI automation team.
00:05:22.800 --> 00:05:28.319
Uh I've been dabbling with AI stuff since co-pilots first started telling me what to do.
00:05:28.480 --> 00:05:39.439
Uh tried to make that into a predictive text generator for functions, and ever since I've been sort of captivated with making AI a productive tool for my development workflows.
00:05:39.600 --> 00:05:46.639
So I'm here to pick up on what the current trends and you know feels are in the industry among everyone that's here.
00:05:46.879 --> 00:05:57.279
Amazing, and of course, we first met, uh Alan Pope introduced us, and we first met at um at Starbucks like 18 months ago, and other other bad coffee establishments are available.
00:05:57.360 --> 00:06:04.879
Uh but you know, from when we first chatted to now, what would you say are some of the biggest changes in your working environment and style?
00:06:05.199 --> 00:06:06.639
Well, agentic development point.
00:06:06.959 --> 00:06:15.600
When we spoke back then, we were talking about the capability of the models, and then shortly after that, the whole sort of MCP and agentic thing kicked off.
00:06:15.839 --> 00:06:19.439
So I think it's really the tooling around the models that has changed the most.
00:06:19.680 --> 00:06:26.959
Obviously, the model capability has improved significantly, but I think the force multiplier has been the tooling around the models.
00:06:27.120 --> 00:06:37.600
So, skills, MCP servers, they've really sort of delivered sort of freesency of information to the models that really make them more much more capable.
00:06:38.319 --> 00:06:40.319
So I think that's the the big thing for me.
00:06:40.560 --> 00:06:45.120
And uh and other than my conference opening, what would you say is your highlight for the uh the sessions?
00:06:45.360 --> 00:06:50.399
Uh the sessions, um I don't know, that's a tough one.
00:06:50.560 --> 00:06:53.680
I think um there's been a couple of the exhibitors that are interesting.
00:06:53.839 --> 00:06:57.199
So uh the uh virtual SRE I really like.
00:06:57.279 --> 00:07:07.600
The uh looks inside your existing observability stack and acts like a historical SRE looking at where the issues were and automatically proposing fixes.
00:07:07.680 --> 00:07:08.560
So I kinda like that.
00:07:08.639 --> 00:07:12.079
Rather than which one it's uh is it autonomy AI?
00:07:12.160 --> 00:07:12.639
Is that the one?
00:07:12.800 --> 00:07:13.680
Yeah, autonomy AI, yeah.
00:07:13.839 --> 00:07:14.079
Yeah, yeah.
00:07:14.399 --> 00:07:14.639
Awesome.
00:07:14.800 --> 00:07:15.680
We'll talk to them in a bit.
00:07:15.839 --> 00:07:17.120
Martin, thank you for spending the time with us.
00:07:17.279 --> 00:07:17.680
Cheers.
00:07:21.199 --> 00:07:22.639
What's your name and what do you do?
00:07:22.800 --> 00:07:24.800
My name's Ryan, I'm uh QA team lead.
00:07:25.040 --> 00:07:25.680
QA team lead?
00:07:25.920 --> 00:07:26.160
Very interesting.
00:07:26.480 --> 00:07:31.360
What are what are the things that you've come to the R Native DevCon to learn or practice more about?
00:07:31.519 --> 00:07:38.720
Yeah, I've come to uh potentially see gaps, like from what I know and what I don't know, and I've identified a few gaps.
00:07:38.959 --> 00:07:39.199
That's great.
00:07:39.519 --> 00:07:39.680
Awesome.
00:07:39.759 --> 00:07:40.959
Would you mind sharing a couple of those with us?
00:07:41.199 --> 00:07:45.279
Well, sort of harnessing or harnesses as a whole, yeah, that's a new concept to me.
00:07:45.600 --> 00:07:47.600
And uh evals
is something I'm very interested in.
00:07:47.759 --> 00:07:53.920
Evals for your old context and what are the what are the kind of things that you feel like you can take away and try straight away?
00:07:54.160 --> 00:07:55.439
The evals for sure, yeah.
00:07:55.600 --> 00:08:01.040
We've got skills spread all over the place and kind of want to consolidate and look into the evals.
00:08:01.360 --> 00:08:03.680
What would you say are some of the biggest problems around skills?
00:08:03.759 --> 00:08:06.399
Is it is it distribution, is it duping of skills?
00:08:06.720 --> 00:08:17.360
A bit of duplication, um, but I'd say at work, it's just skills coming in too quickly and more on the mid to junior levels, they're not knowing about the skills and how to use it.
00:08:17.439 --> 00:08:19.839
So it's a bit of it's consolidation in my opinion.
00:08:20.399 --> 00:08:22.639
Consolidation, and I guess it sounds like a little bit of education about the.
00:08:23.199 --> 00:08:24.399
Education is that that's correct.
00:08:24.800 --> 00:08:25.360
Amazing, amazing.
00:08:25.519 --> 00:08:26.560
Anything else you're looking forward to?
00:08:26.639 --> 00:08:28.959
Anything else you got your name by in the in the workshops or sessions?
00:08:29.199 --> 00:08:32.720
Next workshop, yes, it's the uh tests, uh what's it?
00:08:33.120 --> 00:08:35.759
Yeah, it's a uh I can't remember the name of it now, yeah.
00:08:35.840 --> 00:08:38.080
But it's a test um test strategy, I think.
00:08:38.240 --> 00:08:38.720
Oh, awesome.
00:08:39.200 --> 00:08:42.240
Yeah, yeah, it's that one with Baruch and uh Macy maybe perhaps from Terrace.
00:08:42.480 --> 00:08:44.399
Yes, awesome, it's gonna be a great workshop.
00:08:44.639 --> 00:08:44.879
Excellent.
00:08:45.039 --> 00:08:46.960
Thank you very much for speaking with joining the rest of the conference.
00:08:50.480 --> 00:08:54.240
But we have uh Manny Sarkar here, uh long-term supporter of the INA tier.
00:08:54.320 --> 00:08:56.639
David, it's great to see you here at the conference as well.
00:08:57.039 --> 00:09:02.399
So, Manny, tell us a little bit about your background, uh, tell us a little bit about why you're here, what are you trying to learn at the conference?
00:09:02.559 --> 00:09:03.120
Yes, definitely.
00:09:03.200 --> 00:09:07.279
Uh so I'm uh I'm a software engineer turned AI slash ML engineer.
00:09:07.519 --> 00:09:22.559
I've got a history of 25 years of software engineering, uh and I in the last eight to ten years, I switched to or I started taking on data science and machine learning a lot more seriously, and I leverage my software engineering experience to learn and implement all of that.
00:09:22.720 --> 00:09:31.919
And so today what I do is I if I solve a machine learning or an AI problem, I wear my software engineering, my best practices hat, and I solve it in from that perspective.
00:09:32.240 --> 00:09:32.559
Interesting.
00:09:32.720 --> 00:09:43.279
What would you say are some of the best practices that you feel uh an agentic coder of today could very easily drop or miss when they're when they're trying to build out their agentic uh workflows and processes?
00:09:43.519 --> 00:09:46.000
Yeah, so uh I was just talking to one of the attendees here.
00:09:46.559 --> 00:09:47.200
I just really insert.
00:09:47.600 --> 00:09:50.639
Yeah, yeah, no, that's okay, but I think it segfits very well into what we're saying.
00:09:50.799 --> 00:09:58.720
So what's happening in ST's days is we are reading less and we are passing on the buck over to the other side, which is non-deterministic.
00:09:58.799 --> 00:09:58.879
Yeah.
00:09:59.120 --> 00:10:08.399
We need to build and discipline ourselves to to with deterministic processes and steps and try to reduce the non-determinism.
00:10:08.639 --> 00:10:17.120
So uh using a whole swarm of agents to build something, but then not knowing what it has built, so you don't know how to use it, and then uh that's where it's one of the things.
00:10:17.200 --> 00:10:24.879
The other bit is like not having a plan and just like you know, vibe coding itself is like just writing prompts after prompts, isn't the greatest practice.
00:10:25.120 --> 00:10:26.240
You plan your work.
00:10:26.320 --> 00:10:40.080
So, like I was telling the uh the attendee that I remember Einstein's quote, he says that if you give me a problem and you give me an hour, I'll spend 55 minutes trying to think about the problem, trying to brainstorm the problem, and I'll spend five minutes trying to solve the problem.
00:10:40.159 --> 00:10:44.559
Now, Einstein, great mind, has done some great work or lots of great work.
00:10:44.720 --> 00:10:47.840
We can learn something from that, and then we can apply it in our space.
00:10:48.000 --> 00:10:48.559
Yeah, yeah, yeah.
00:10:48.639 --> 00:10:49.039
Amazing.
00:10:49.200 --> 00:10:53.679
What is some of the so what what for example were some of the sessions that you really took things uh good good practice?
00:10:53.919 --> 00:10:56.080
So Guy's session kicked off really well.
00:10:56.159 --> 00:11:00.159
I I love, and I'm gonna randomly name them, but Patrick's session was amazing.
00:11:00.240 --> 00:11:07.519
Uh I would love to work with someone like Patrick who knows how he needs to run, how teams need to be run while we are in the AI space.
00:11:07.679 --> 00:11:11.279
Uh uh Dave Farley really closed it up well yesterday.
00:11:11.759 --> 00:11:23.759
Um and then there are some other sessions where I took little bits of uh things like um pass off, do treat your current practices just like you were before AI came in.
00:11:24.080 --> 00:11:25.759
Don't don't drop those practices.
00:11:25.919 --> 00:11:33.919
What what we've been given is AI to augment our skills and not to pass it on to AI to do all of the stuff and then not know what's happening.
00:11:34.399 --> 00:11:51.519
Because it's a great steer, it's a great skill, a great way to make us 10x, but we need to be
constantly disciplined, have a have a plan, uh have a structure, uh, and use it in that manner and not slip away by a lot of people who are doing by coding billions of lines of code and not knowing actually what it has created.
00:11:51.759 --> 00:11:52.720
AI slot, basically.
00:11:53.039 --> 00:11:53.120
Yeah.
00:11:53.360 --> 00:11:53.600
Amazing.
00:11:53.679 --> 00:11:55.840
Well, Manny, thank you very much for your time.
00:11:58.799 --> 00:12:01.279
So, Kid, you're the founder of Near Forms.
00:12:01.360 --> 00:12:01.440
Yes.
00:12:01.600 --> 00:12:06.320
Uh tell us a little bit about what you're learning here from attendees, some of the conversations that you've had at the conference.
00:12:06.480 --> 00:12:07.279
Oh my god.
00:12:07.440 --> 00:12:13.440
Um I've I've learned a lot about uh uh harness engineering while I've been here.
00:12:13.519 --> 00:12:19.840
It's it's crazy from when the conference was announced to now, like it's become a thing, like and it you know, it wasn't even a thing like three months ago.
00:12:20.000 --> 00:12:30.799
This is the horrible thing about a CFP for a conference like this, because you open a CFP and then something that happened, you know, we opened it three, four months ago, and the content and the things that are gonna be hot, you don't know what's gonna be hot by the time the conference comes around.
00:12:31.120 --> 00:12:32.080
So harness engineering.
00:12:32.320 --> 00:12:34.240
Harness engineering uh tokenomics.
00:12:34.399 --> 00:12:34.639
Yeah.
00:12:35.279 --> 00:12:39.679
Everyone's going like, how are we gonna afford to favor this stuff in a few months' time, you know?
00:12:39.919 --> 00:12:43.120
Um, and uh just a really vibrant community here, you know.
00:12:43.200 --> 00:12:43.519
Yeah.
00:12:43.840 --> 00:12:55.679
Um so uh an uniform, obviously, a consultancy that come and help people in a very agentic way, which is which is super super cool for a consultancy to really be uh to really be pushing that.
00:12:56.000 --> 00:13:01.919
Tell us a little bit about what some of the problems, some of the some of the things that your clients and customers are uh are asking for you these days.
00:13:02.159 --> 00:13:08.399
Yeah, um we're do we're doing a combination of helping uh companies to uh start doing things in an agentic way.
00:13:08.480 --> 00:13:17.279
Uh you know, you have in most companies like one or two teams that are doing things in a more advanced way, and it's like, but how can I get to 100 teams so we're we're solving that problem for them?
00:13:17.440 --> 00:13:31.039
Uh so we're you know we're coming in, we're doing kind of hands-on training workshops, uh, building with the customers, so we technically uh you know, pick a couple of lighthouse projects, work with our teams, and get them into kind of higher confidence level of what they're doing.
00:13:31.200 --> 00:13:33.519
Um just doing a bunch of stuff like that.
00:13:33.679 --> 00:13:36.399
Um and then we're doing some applied AI stuff as well.
00:13:36.559 --> 00:13:43.039
So uh we have an interesting use case um for like a uh an AML project with a financial institution.
00:13:43.519 --> 00:13:50.000
Um so the AML department are uh you know like uh creaking at the seams to do like KYC stuff and all that.
00:13:50.080 --> 00:13:53.200
Like so I've done an agent to do that stuff recently.
00:13:53.440 --> 00:13:58.159
Um so uh I think we I think we cleared like a thousand AML cases in 20 minutes.
00:13:58.320 --> 00:13:58.559
Wow.
00:13:58.720 --> 00:14:03.200
Uh whereas that would have been about three to four thousand hours of human time to do it, like you know.
00:14:04.000 --> 00:14:13.919
Amazing opportunities for for kind of like you know, scaling and uh being uh being more efficient and effective with uh yeah, yeah, the other stuff we're seeing is um around modernization.
00:14:14.480 --> 00:14:43.360
Um so like uh we've a fun one with a client uh where where a couple of years ago we were we budgeted like um a modernization project to be like uh it was over 10 million uh euro project, like big old code base, uh been around a long time, and uh you know we so we kind of spec it up free AI and recently we we've kind of looked at doing it and it's it's coming in well under two million to do the same project in about about uh about 35% of the time.
00:14:43.519 --> 00:14:54.559
So like you've seen this like massive acceleration uh in in in uh uh the the or reduction in the amount of time it takes to do something, yeah, yeah.
00:14:54.720 --> 00:15:00.879
You know, um so like we're seeing it from a consulting perspective, like you know, this this huge opportunity, you know.
00:15:01.039 --> 00:15:01.519
Yeah, yeah.
00:15:01.600 --> 00:15:02.000
It's amazing.
00:15:02.240 --> 00:15:04.240
How have you found the crowd here at AN Native DEF CON?
00:15:04.480 --> 00:15:06.159
Oh great, yeah, yeah, yeah, yeah.
00:15:06.559 --> 00:15:08.720
Lovely community, like very chill out for us.
00:15:08.960 --> 00:15:09.279
Yeah.
00:15:09.440 --> 00:15:11.759
Um yeah, people are very nice here, you know.
00:15:12.000 --> 00:15:12.399
Amazing.
00:15:12.480 --> 00:15:14.240
Um so really enjoyed it, like, yeah.
00:15:14.480 --> 00:15:14.799
Wonderful.
00:15:14.879 --> 00:15:18.000
Well, thank you very much and thank you for your all your support for Air Native DEF CON.
00:15:21.120 --> 00:15:26.000
I think it's a really interesting diversity in how where people are at with their adoption of AI, yeah.
00:15:26.240 --> 00:15:31.759
Um from the kind of cautious but curious end of the spectrum through to the kind of YOLOing it.
00:15:32.159 --> 00:15:35.360
And I think my take on that is there's a lot of wisdom in both ends.
00:15:35.519 --> 00:15:46.080
Um, people are rightly cautious with wisdom that's been hard won over the years, and they're looking for something that helps them to ensure that AI stays on the right track.
00:15:46.159 --> 00:15:54.000
And but I'd be able to
unleash it, unleash the speed of AI without just being kind of letting it do whatever and and just accepting the risk.
00:15:54.080 --> 00:15:54.240
Yeah.
00:15:54.399 --> 00:15:57.360
And what do you say are some of the, in terms of the, sorry, I'm I'm good.
00:15:58.080 --> 00:16:09.840
Yeah, what would you say are some of the some of the biggest problems that people have in their organizations today where you know a mature security solution, which is very AI-first, can can kind of like help you for them?
00:16:11.759 --> 00:16:13.279
A lot of it comes down to trust.
00:16:13.440 --> 00:16:13.759
Yeah.
00:16:14.000 --> 00:16:23.919
Um and how you can uh trust that things are gonna stay on track without having to be hands-on and watching everything.
00:16:24.399 --> 00:16:38.399
And so I think where Sneak has a lot to offer there is providing the trust, is giving a solution that people can rely on Sneak to ensure things don't go wrong, with and and don't feel like they have to be hands-on, you know, gripping the wheel really tightly.
00:16:39.519 --> 00:16:42.080
At scale, I presume, as well, across organ across the organization.
00:16:42.320 --> 00:16:42.799
It's got to scale.
00:16:42.879 --> 00:16:42.960
Yeah.
00:16:43.039 --> 00:16:48.320
It's got to scale, it's got to be codified, um, it's got to be something that can fit into workflow again without slowing things down.
00:16:48.399 --> 00:16:48.480
Yeah.
00:16:48.639 --> 00:16:50.960
Um and actually integrated in a way that flows.
00:16:51.440 --> 00:16:51.759
Awesome.
00:16:51.919 --> 00:16:55.759
Great to see Sneak here, great to see him supporting the uh the native dev community server.
00:16:56.080 --> 00:16:56.480
It's been really fun.
00:16:56.720 --> 00:16:57.120
It's been really fun.
00:16:57.279 --> 00:16:58.639
Amazing, thank you very much.
00:17:02.080 --> 00:17:04.079
So, Adir, tell us a little bit about Autonomy.
00:17:04.160 --> 00:17:05.599
You're the founder and CEO, I believe.
00:17:05.759 --> 00:17:06.079
CEO?
00:17:06.240 --> 00:17:06.319
Yes.
00:17:06.480 --> 00:17:06.960
Yes, there we go.
00:17:07.039 --> 00:17:08.160
I didn't just uh promote you, did I?
00:17:08.240 --> 00:17:08.400
No.
00:17:08.559 --> 00:17:10.400
The founder and CEO of Autonomy AI.
00:17:10.559 --> 00:17:12.000
Tell us what is Autonomy AI?
00:17:12.079 --> 00:17:15.839
So Autonomy AI is basically an operating system on existing code bases.
00:17:16.079 --> 00:17:30.720
We enable, we understand existing code bases, understand the components, the design system, all the design library, and then we enable non-technical or semi-technical individuals, think about your PMs, your designers, your professional services, to be part of the delivery process.
00:17:30.880 --> 00:17:35.200
Instead of going back and forth with engineering, they can deliver everything end-to-end on your production code.
00:17:35.440 --> 00:17:47.200
And what would you say are kind of like some of the biggest differences between a non-technical person and a technical person when it comes to them uh doing work in the agentic workflow, raising PRs, all that kind of thing.
00:17:48.079 --> 00:17:49.119
What are the differences in needs?
00:17:49.359 --> 00:17:51.519
Well, let's start from the basic, the interface.
00:17:51.759 --> 00:17:56.160
All the workflow or all the interfaces you have today are usually CLI interfaces, right?
00:17:56.400 --> 00:17:59.599
And today what we enable is a web interface and it's super easy.
00:17:59.759 --> 00:18:00.480
That's the first thing.
00:18:00.640 --> 00:18:10.240
The second thing, in order to actually push a PR that is valid, viable, in a consistent, reliable way, uh, the PMs don't know what the cons what are the constraints of the platform itself.
00:18:10.400 --> 00:18:17.680
What autonomy can help them with and what Face Studio our product is helping them with, is just click a button, the pull request will understand the entire system.
00:18:17.839 --> 00:18:24.480
PMs need to know how to ask visual things, uh uh KPIs on the product, not how to build the code behind the scenes.
00:18:24.720 --> 00:18:25.440
Super interesting.
00:18:25.519 --> 00:18:29.839
Uh of course we have Tomaz, uh your CTO here uh in doing one of the sessions.
00:18:29.920 --> 00:18:42.960
Uh what when you're talking with some of the folks here at the booth or in sessions or on the hallway track, uh what are what are people's biggest problems today that that they're that they're folks that they're facing with their agentic coding uh movement?
00:18:43.279 --> 00:18:49.359
Everyone is spending so much money, so much effort on this, but not actually seeing any needle movement.
00:18:49.759 --> 00:18:54.000
The needle movement is very, very small, around the 5 to 10%.
00:18:54.559 --> 00:18:55.680
And that's the main challenge.
00:18:55.759 --> 00:19:01.759
And when we talk with people, we don't actually just talk about our product, we talk about the actual workflow, where are the real bottlenecks?
00:19:02.000 --> 00:19:10.160
When we start exposing this, they understand that the bottlenecks are not, you know, the coding velocity, writing code, it's actually the workflow that you have within the organization.
00:19:10.319 --> 00:19:15.680
And in one word, what is the what is the thing that people should focus on that they're not doing spending enough time on?
00:19:15.920 --> 00:19:16.720
Change management.
00:19:16.960 --> 00:19:22.000
Utilize those tools, not just to write things faster, but actually how you can orchestrate your organization faster.
00:19:22.480 --> 00:19:24.240
Change management, and we'll call that with a hyphen.
00:19:28.240 --> 00:19:35.599
What are the biggest problems that people in organizations are having today in and around their agentic coding journey?
00:19:35.920 --> 00:19:37.359
Who wants to take that one?
00:19:38.799 --> 00:19:49.920
So we have been talking to a bunch of people who are getting very excited about agents and
then installing them all over the place and making lots and lots of skills, and then they're a bit nervous about all the skills.
00:19:50.079 --> 00:19:59.200
So they're either worried about security or they're worried about how they keep track of all of the different ones and work out how to share them properly and not lose track.
00:19:59.440 --> 00:20:01.200
So I've been hearing that a lot from different people.
00:20:01.759 --> 00:20:08.480
So sharing them through distribution, being able to actually validate them and understand if they're good or if they're secure, those types of things.
00:20:09.440 --> 00:20:10.640
Uh have you been to many sessions?
00:20:10.880 --> 00:20:12.720
I'm excited to go to my first session today.
00:20:12.960 --> 00:20:13.920
Oh, which session are you going to do?
00:20:14.079 --> 00:20:17.200
I'm seeing Don Sim talk on automation in GitHub.
00:20:17.519 --> 00:20:23.519
So I'm really looking forward to that because I had a chat with him yesterday and he kind of like gave me gave me the gist.
00:20:23.599 --> 00:20:30.880
And I think he has a point about GitHub and GitHub actions actually being a big part of automation in the future.
00:20:31.200 --> 00:20:33.519
So I'm really curious to see what else he has to say about that.
00:20:34.160 --> 00:20:41.519
How ready is is Git and GitHub, would you say, for uh an agentic coding workflow of today?
00:20:42.400 --> 00:20:45.519
Let me put this a let me put this questionable question to you.
00:20:45.839 --> 00:20:46.720
A contentious question.
00:20:47.359 --> 00:20:51.440
I think they have the opportunity, but being able to seize that is different.
00:20:52.000 --> 00:20:59.839
What I mean by that is like when you've got the core primitives of like an action is just basically running a workflow in the cloud that the user is able to define.
00:21:00.079 --> 00:21:07.279
And so that as a core primitive feels very valuable when you've got an agent that you can now run as part of that workflow in terms of the sequencing, in terms of forking, in terms of branching.
00:21:07.759 --> 00:21:15.839
But uh to be able to capitalize on that, they need to make sure that it's reliable and it is effective, and that they've built the interfaces for people to do that.
00:21:16.079 --> 00:21:21.680
Um, particularly given it code is now so cheap that people can go build build their own versions themselves.
00:21:21.920 --> 00:21:34.559
And so having the ease of onboarding so much so that it is easier for me to get an agent to use GitHub actions rather than for me to go have an agent write its own orchestrator on my own system, I think is kind of the important tension point there.
00:21:35.519 --> 00:21:44.799
And we'll wrap up with uh one final question about uh a very interesting early access uh that Tesla announced here at GIFO's session in and around an agent harness.
00:21:45.039 --> 00:21:47.680
Uh tell us a little bit about what that is, what people can expect.
00:21:48.240 --> 00:21:57.359
Yeah, so for Tesla agent, I think the main framing is people are gonna need to shift up the abstraction sack to be doing harness engineering rather than just telling an agent what to do.
00:21:57.519 --> 00:22:02.079
And to do that effectively, you need to set up your environment so that the agent can run effectively in that.
00:22:02.160 --> 00:22:10.240
You've got hooks, you've got contacts, you've got verifiers, you've got all these kind of core concepts that you can use to make sure the agent is producing the correct code.
00:22:10.400 --> 00:22:17.839
The Tesla agent is good is there to help you set up that environment and set up your ecosystem and set up your code base for a genetic success in it.
00:22:18.000 --> 00:22:31.359
So it shifts your human leverage point to the point where you have the most value in terms of defining how the agent should be operating rather than telling the agent what it should be doing, because that's something that can be much tighter, uh much more easily encapsulated in a ticket.
00:22:31.680 --> 00:22:31.920
Amazing.
00:22:32.000 --> 00:22:37.839
And if you're interested in that, TESL.co forward slash agent, and you can learn more about how to get onto the early access list.
00:22:41.440 --> 00:22:45.680
So that's a wrap for IO Native DevCon 2026 in London.
00:22:45.839 --> 00:22:54.640
I'm here in the VIP room where we essentially let speakers and and uh some of our VIP attendees hang out, grab good coffee in nice comfort.
00:22:54.880 --> 00:22:56.000
I'm pretty tired now.
00:22:56.160 --> 00:23:04.000
My mind is uh absolutely full of amazing information from tons of great speakers, tons of great conversations in the in the in the hallway track.
00:23:04.240 --> 00:23:09.680
I think it's time to maybe have a little rest and then uh and start putting all that great knowledge into practice from tomorrow.
00:23:10.000 --> 00:23:15.200
Thanks all for watching, thanks for attending, and uh hopefully we'll see you in New York or London next year.
00:23:15.680 --> 00:23:16.559
Bye for now.
00:23:16.880 --> 00:23:21.680
The AI Native Dev is brought to you by Tesla, the package manager for Skills and Compact.
00:23:21.920 --> 00:23:24.880
Your hosts are Guy Pajani and me, Simon Maple.
00:23:25.039 --> 00:23:26.559
Our producer is Tom Dowler.
00:23:26.799 --> 00:23:29.920
The AI Native Dev is not just a podcast, it's a community.
00:23:30.160 --> 00:23:33.599
And we host monthly meetups at the Tesla offices in central London.
00:23:33.839 --> 00:23:36.240
Visit Tesla.io forward slash community.
00:23:36.720 --> 00:23:38.880
To learn more, I hope to see you there.