Nov 29, 2024 · 48m · news
Guy Podjarny: The Future of AI Software Development - What is Real & What is BS | E1232 · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 20VC, Harry Stebbings interviews Tessl and Snyk founder Guy Podjarny to dissect the realities of AI software development, debunking SaaS replication myths, examining the changing role of developers, and sharing critical fundraising and operational strategies for founders.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 20.2% of the talking time here. How this is scored →
Harry as informed peer 4.2 Guest teaching 5.0 Guest disagreement 2.8 Harry pushing back 4.0
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 13:22 Guy rejects AI SaaS replication as bullshit
Guy emphatically dismisses the premise that AI tools can easily recreate existing SaaS companies, calling the notion 'bullshit' and arguing that SaaS value lies in data, distribution, and switching costs.
Hardest push from Harry ▶ 25:31 Harry rejects Guy's defense of Google SearchHarry forcefully refuses to accept Guy's argument that Google will maintain search dominance, arguing directly that Google Search product quality sucks compared to Perplexity.
Biggest teaching moment ▶ 43:38 Guy lectures on the traps of overfunded startupsGuy provides a masterclass on capital allocation mistakes, breaking down how large balance sheets destroy founder urgency and cause companies to scale prematurely before reaching true product-market fit.
Harry holds his own ▶ 6:18 Harry links AGI timeline claims to fundraising incentivesHarry demonstrates sharp venture capital expertise by contrasting short-term AGI hype from founders who constantly need to raise capital against longer-term timelines cited by well-funded leaders.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome & Quick-Fire: Is Nvidia Undervalued? | 4 | 4 | 2 | 5 | Harry challenges the assumption of endless chip demand by pointing to a potential trough of disillusionment and AI ROI shortfalls. Guy breaks down Nvidia's durability into three distinct strategic layers. | |
| AGI Timeline, Capex, and Fundraising Incentives | 5 | 4 | 2 | 3 | Harry introduces an insightful observation from Tom Hume linking aggressive AGI timelines directly to fundraising incentives. Guy agrees while expanding on non-technical societal barriers to AI adoption. | |
| Scaling Laws, Specialized Models, and Tessl's Strategy | 3 | 5 | 2 | 2 | Guy educates Harry on specialized models versus general foundation models, positioning Tessl as a user of cloud infrastructure rather than a model trainer. Harry synthesizes the trade-off clearly. | |
| Evaluating AI Dev Tools: Copilot, Cursor, and the Jagged Edge | 4 | 6 | 3 | 3 | Harry brings quotes from Marc Benioff and Gartner criticising Copilot reliability. Guy explains the concept of the 'jagged edge of AI' where non-human errors viscerally frustrate users. | |
| Why AI Cannot Easily Replicate SaaS Businesses | 3 | 7 | 7 | 4 | Harry floats the popular claim that AI tools can easily replicate complex SaaS products like ServiceTitan. Guy strongly rejects the premise as 'bullshit', lecturing on defensibility beyond pure code. | |
| Fundraising Dynamics and the Threat of Closed AI Ecosystems | 5 | 5 | 4 | 6 | Harry questions if Tessl's massive round is VC excess and pushes back on Guy's fear of closed AI ecosystems by highlighting OpenAI's platform model. Guy defends Tessl's reserves and maintains his warning about locked gardens. | |
| Platform Monopolies and Enterprise Budget Shifts | 4 | 5 | 2 | 2 | Harry tests Guy with a structured multi-choice valuation scenario across frontier model startups. Guy picks Anthropic based on team stability and breaks down enterprise software budget migrations. | |
| OpenAI's Leadership Crisis and Google's Unbeatable Search Moat | 6 | 5 | 5 | 8 | Harry forcefully pushes back when Guy defends Google's search moat against Perplexity, arguing directly that Google Search Sucks in comparison. Guy holds his ground citing habit inertia and distribution. | |
| Trump's Regulatory Impact, Snyk's IPO Outlook, and Missing the CEO Role | 4 | 4 | 2 | 3 | Harry questions why mature tech companies still need to IPO given abundant private capital. Guy explains enterprise credibility, brand recognition, and equity liquidity benefits. | |
| Managing Architectural Overlap and the Rise of Developer-PMs | 5 | 6 | 3 | 6 | Harry presses on the potential chaos if every developer evolves into an architect without clear leadership structure. Guy counters by highlighting the vast expansion of software volume needing architectural trade-offs. | |
| Tech Team Structures and the Hidden Security Risks of AI Code Rot | 4 | 6 | 3 | 4 | Harry banteringly teases Guy for making VC-like sweeping predictions before redirecting to code security. Guy educates on code rot and unmaintained AI-generated code risks. | |
| Spicy Q&A: Secondary Liquidity, Acquisition Offers, and Worst VC Meetings | 3 | 2 | 1 | 5 | Harry conducts a high-stakes charity Q&A asking direct financial and personal questions. Guy surprises Harry by transparently revealing nine-figure secondary liquidity and worst VC experiences. | |
| AI Scaling Expectations and Ignored Opportunities | 5 | 5 | 3 | 4 | Harry brings up an internal VC debate regarding whether crowded markets validate demand. Guy advises founders to seek ideas requiring active conviction rather than crowded 50-competitor spaces. | |
| Fundraising and Cash Management Strategy for Tessl | 5 | 7 | 2 | 3 | Harry calculates interest earnings on Tessl's $125M balance sheet. Guy delivers a detailed mini-lecture on the two critical traps overfunded founders fall into regarding burn rate and lost urgency. | |
| Guy Podjarny's Best Angel Investments | 3 | 4 | 1 | 2 | Harry wraps up with casual questions on angel investments and marriage. Guy shares his top angel investments and his rule on immediate marital communication. |
Statements from this episode (48)
Podjarny: SaaS companies differentiated only by code are doomed
“SaaS businesses are far more than just the software that they create. In fact, if you have a SaaS business and your only differentiation is I've written all this code and nobody else can do, then your days are numbered.”
Podjarny: Architectural issues could stall GPT-5 scale models for 1-2 years
“Expectation of when would the GPT-V scale models arrive? I now think that those are further away. You've sort of seen Sam Altman and others change their tune to talk about reasoning and things like that as the progress, and my sense is there are actually some …”
Guy Podjarny: Nvidia is building a cloud to leverage chip lead
“They know that the clouds, for instance, have distribution advantages, and so they're building a cloud. They're using their semiconductor advantage.”
Podjarny: Nvidia's revenue is largely secured by existing advance commitments
“They sold a lot of commitments already, and so I think a good amount of their revenue is pretty guaranteed”
Podjarny: Many AI enterprise deployments will go sideways due to slow organizational adaptation
“Well, I think a lot of AI budgets at the moment are very kind of non-resilient, right? They're coming in, people are spending money, they're trying things out, and they expect a lot from them, and I think in the long run that would be correct. But it's hard to…”
Podjarny: Massive capital waste occurring in AI due to thousands of duplicate startups
“I think the biggest one also is all of these tiny startups where there's like a thousand, I think in some cases there might be even 10,000 companies doing the exact same thing. So it almost doesn't matter how good they are and how good the use case is. A lot o…”
Podjarny: Legal and insurance issues will delay AI adoption more than tech
“I think people underestimate the non-technology parts of adopting AI. I think AI is a transformative or a disruptive technology, not a sustaining one, and it requires people to change, it requires kind of accountability to change, insurance, and things like th…”
Podjarny: Near-term AGI predictions are convenient because AGI lacks clear definition
“It's a very convenient thing to say because AGI is very poorly defined, and so you can basically say that and pretty much claim it is correct.”
Podjarny: Near-term AGI claims correlate with founders' continuous need to raise capital
“I think that's absolutely true. And there's incentives. Incentives are not always nefarious. Sometimes it's just your view of the world is biased by the things that you want to believe.”
Podjarny: Specialized AI models will prove out over next three years
“Yeah, I think I am more willing to bet that the specialized models will kind of prove out in the sort of, call it three year timeline than I did before. Just because I think the big models, the big generic models are apparently, and this is again a bit of rumo…”
Podjarny: Tessl will not train foundation models or lock into one vendor
“You know, I don't want to, I definitely don't want to compete with all of those, but I also don't even want to pick amidst them. I want to be able to use The best technology available to me at any time. If you draw the cloud analogy, I want to be the best user…”
Podjarny: AI coding tools lower code quality by turning developers into reviewers
“I do think that it develops software of questionable quality in many places, because You're just kind of reviewing the codes that got generated as opposed to actually writing it, and so just naturally the amount of thought you put into it is, is, is far lesser…”
Podjarny: Easiest AI tools fit workflows without needing complete user trust
“When you think about AI solutions as a whole, if I think kind of broader mental framework, the solutions that are easiest to adopt right now are the ones that really don't require any change in how you work today, and they just provide you with some magic, and…”
Podjarny: AI developer tools remain overall unreliable due to unpredictable failures
“Beyond that, they haven't been dramatic yet, mostly because they are still unreliable. They're you know, we our founding AI engineer calls it the jagged edge of AI. You know, it is amazing at some point, and then it bombs, like, a complete disaster the second.…”
Podjarny: Replicating SaaS businesses via AI agents is silly
“To think that you can just replicate a SaaS business because you can tell an agent what it would do, I think is I think is silly.”
Podjarny: Massive gap exists between impressive AI demos and working products
“It's very easy to create amazing videos in AI, amazing demos and it's, there's a big, big gap between that and creating a reliably working product, hence the jagged edge of AI.”
Podjarny: Everyone who tested Cognition's Devin says it doesn't really work
“From everybody that I talked to that has tried Devon, they all think it's really cool and that it doesn't really work.”
Podjarny: Tessl is not targeting a quick $1B-$2B exit
“For me, that wasn't an issue, and I am very much sort of not looking to, you know, make some sort of quick one, two billion dollar exit.”
Podjarny: AI platform concentration risks marginalizing developer tooling market
“I worry that we go into a place in which there are a small number of players that have that broad capacity and that the rest of the dev tooling ecosystem becomes kind of minor and delegated versus today's world where we have many thriving developer tooling com…”
Podjarny: Platforms like GitHub and Vercel will dominate AI software development
“I think there is a pretty high chance that we're going towards this path. Ease of use of software development is already, you know, amazing, and you're seeing platforms like, ah, Vercel, for instance, you know, make it very easy to generate applications and go…”
Podjarny: Autonomous AI agents are years away due to reliability systems
“I think the rest are probably a couple years away, and as I said, I don't think the LLMs are the limiting factor. I think it's more the processes and the systems around them to make the LLMs more reliable.”
Podjarny: Foundational AI models will eventually replace traditional web search
“I think as long as we are dealing with the same data and just a different way to interpret it, I think they will go all the way. And so search is a good example of that, right? There's no reason why they wouldn't go all the way to actually replacing web search…”
Podjarny: AI model providers won't build complex, workflow-heavy software
“I think when it becomes Different data or very dedicated, elaborate workflows. Like, it's not about the magical make a call, get back a response, but a whole system around how do you process those. I think that's outside of their domain and wouldn't make sense…”
Podjarny: Anthropic offers better investment value than OpenAI or xAI
“I think I go Anthropic. I feel generally the, it still holds to the quality of the team or maybe you know, the stability of the team and their opportunity to grow. And yeah, the value arbitrage is is still hard.”
Podjarny: Perplexity won't succeed as expected due to Google's dominant search moat
“I still think perplexity is not going to go as far as people think. Mostly not because of OpenAI, but because of Google.”
Podjarny: Google is terrible at product development except for search
“Google is generally pretty terrible at product on many fronts, but the one product in which they actually do a pretty good job is the search product, and they have all the traffic.”
Podjarny: Google doesn't need to innovate to beat Perplexity, just copy it
“I don't think they have the product jobs, but I think they have the advantage of such, like, incredible distribution and access to the users that they don't actually need to invent what perplexity is doing. They can copy it.”
Guy Podjarny: Mid-tier tech companies will acquire easier under Trump
“The very top layer, the top companies that Trump affiliates as liberal and anti-Trump will not have kind of a better life because of his administration, and the companies that are just the layer below, they will be able to acquire a lot more easily.”
Guy Podjarny: Top software developers excel through systems thinking, not coding
“Well, I think the best software developers are not the best because they're the best coders. It's because they think about development as a whole. They are systems thinkers.”
Guy Podjarny: Manual coding will become an edge case within 10 years
“And so I think the coding piece of software developers' work will diminish substantially. I think in 10 years' time, coding would be, ah, ah, very much alive and well, but it would be the edge case. It would be the thing that you do because, you know, like, it…”
Podjarny: Higher developer leverage will drive up software quality expectations
“Developers will have more leverage in their ability to produce high quality software. And consumer expectations, business expectations, will increase accordingly.”
Podjarny: Technical product manager and software developer roles will merge
“Some PMs are quite technical, and they're close to the software. I think those roles will definitely merge.”
Podjarny: AI will alter job scopes, not overall tech company structures
“I don't think a lot of the structure will change dramatically. I think the roles of the different positions, the sort of the scope of what they can do are the things that will change.”
Podjarny: AI tools increase security risks via unreviewed code deployment
“The real challenge with LLMs right now is that we lose all control. We make a request, and we get something back, and we hope that it is correct, and we're reliant on a very, kind of, unreliable review of what was produced to say that it's correct, and that's …”
Podjarny: Unmaintained AI code will rot and create cyber attack entry points
“The other thing that's happening is people produce software, like a lot of these coding systems, they create code, but they don't maintain code, and so that code will stick around. You know, technology never dies. It'll, you know, be another entry point for At…”
Guy Podjarny took out nine figures in secondary liquidity from Snyk
“I sold about a third of my ownership, and it's in nine figures.”
Snyk's last concrete acquisition offer was around $200 million
“The last acquisition that really I sort of felt like we were talking in, in concrete numbers was early in the journey. It was in sort of two hundred million range, and I think later on we talked about multiple billions, but we would, they didn't get to, like, …”
Podjarny: AI builders over-index on short-term optimization vs first principles
“I think everybody's very focused on the short-term gain on these kind of low friction, you know, just optimize my existing workflow, and people are really not imaginative enough. They're not trying to think about if I stop and think about this domain for first…”
Podjarny: Markets with 5 to 10 competitors are optimal for startups
“I think a market of five or 10 companies is actually pretty good. That's healthy, everybody's pushing along, people are You know, raising awareness for you. I think when it's a market of 50 or a hundred, again, you could be amazing, but it's just everybody get…”
Guy Podjarny: Tessl earns $5M to $10M annually in interest income
“So you definitely get into somewhere between that sort of five and 10, right, if you're sort of massively massively conservative.”
Guy Podjarny: Tessl's interest income fully covers its cash burn
“Which is, you know, definitely sort of at the burn level.”
Podjarny: Startup urgency stems from market timing, not runway limits
“I think eventually urgency comes from market timing, not from runway.”
Podjarny: The Two Main Pitfalls of Raising Too Much Early Capital
“And I think the two primary mistakes that I'm seeing with raising too much money too early Is one is spending it too quickly, and the other is not spending it fast enough.”
Podjarny: Cloudinary angel investment grew from near zero to $4M
“Cloudinary went from effectively zero to about four million dollars worth.”
Podjarny: Invested in SecurityScorecard at a $6M valuation
“Yeah, I invested in security scorecard when they were at six million dollar valuation. That was a pretty good one. Unfortunately, it was a very small check at the time.”
Podjarny: Intercom's pivot to AI agent Finn was bold and correct
“One is, is indeed Intercom's bet on this autonomous agent I thought was very, very wise in Finn. And I think that was a bold and correct move to jump onto that. And I think it required a lot of conviction around around this being the future and then subsequent…”