As technology brings the cost of everything else to 0, psychological costs will predominate.
Reality testing is ultimately unavoidable, of course, but I'd guess most people still lean away from that rather than into it. (Our whole culture is set up that way, and most of us get like two decades of Pavlovian conditioning in that direction.)
Edit: Expanded here: https://nekolucifer.substack.com/p/willingness-to-fail-is-no...
If anything LLM chatbots & synthetic users will make the majority of founders evermore comfortable not testing reality.
The bottleneck was never coding...
You will!
As gp says, there's a big difference between theory and practice here, and a lot of the things we needed when we weren't using LLMs are still needed when we are, but it takes a bit of actual practice to work this out. It's still not at the stage where an Ideas Guy can make a real working product without someone on the team actually knowing how to develop software.
At least in my experience, so far. But the world is changing fast.
So goes the thinking, anyway. It's why my couple decades of experience and I still occasionally get to hear from rando cold recruiters desperate to sell someone a "pivot to AI," probably thinking they can lowball me by holding my mortgage over my head in order to screw three times the work out of me that they'd pay for.
I was in this business too long.
Some that came after might be worthy of the title, but those who claim it for themselves aren't.
LinkedIn influencers
But this I just thought was vacuous. I agree with what you wrote, but more to the point, I didn't find any real advice about how a startup should actually change that passed my sniff test. I left the tech startup world about 2 years ago myself, and I'm glad I did, because I just think there are way fewer differentiable opportunities now. That is, even if I accept what Blank says is true, what are all these 2+ year old startups supposed to do - just create some model wrapper/RAG chatbot product like the million other startups out there?
Even in defense, like the article says, there are now a bajillion drone companies, and it looks like a race to the bottom. The most successful plan at this point just looks like the grifter plan, e.g. getting the current president to tweet out your stock ticker.
I'm honestly curious what folks think are good startup business plans these days. Even startups that looked they were "knock it out of the park" successes like Cursor and Lovable just seem like they have no moat to me - I see very few startups (particularly in the "We're AI for X!" that got a ton of funding in the past two years) with defensible positions.
The much more useful posts are “my team and I are doing X with AI”. Of course, the challenge there is that the ones who are truly getting a competitive edge through AI are usually going to be too busy building to blog about it.
He could have ignored the email or engaged on the topic I introduced. Instead he sent me a wikilink to Autonetics. I was left with the feeling that he had no real interest in the topic he wrote about. It was really no big deal. He is a busy guy and doesn't need to engage with strangers. I never read anything by him again because I was left with the feeling he is just phoning these posts in.
For big complex real world problems, and big complex real worlde codebases, the AIs are helpful but not yet earth shattering. And that helpfulness seems to have plateaued as of late.
I am extremely skeptical of posts like this.
These guys don't get to have it both ways.
So hand-waving about how easy it is to have an MVP in days w/o actually experience in doing that seems ironic.
Now, maybe he's saying this based on companies he's funded who've had great success with what he's saying. But it's curious that the only concrete example of a company mentioned is one that's six years old and not operating like that. And in fact, many of the ways he thinks that company went wrong seem completely unrelated to AI?
> Chris is now starting to raise his first large fundraising round. In looking at his investor deck I realized that while he’s been heads down, the world has changed around him – by a lot. The software moat he built with his 5-year investment in autonomy development is looking less unique every day. Autonomous drones and ground vehicles in Ukraine have spawned 10s, if not 100s, of companies with larger, better funded development teams working on the same problem.
> While Chris has been fighting for adoption for this niche market (one that is ripe for disruption, but the incumbents still control), the market for autonomy in an adjacent market – defense – has boomed. In the last five years VC Investment in defense startups has gone from zero to $20 billion/year. His product would be perfect for contested logistics and medical evacuation. But he had literally no clue these opportunities in the defense market had occurred.
> While there’s still a business to be had (Chris’s team has done amazing system integration with an existing airborne platform that makes his solution different from most), – it’s not the business he started.
"Being heads down without paying enough attention to the market for 6 (!!) years" doesn't seem like an AI-caused issue.
Meanwhile, the core suggestion doesn't seem to fix that, it seems almost completely perpendicular.
> You can now test multiple versions of the same business at once (or simultaneously be testing different businesses). While you can be simultaneously testing five pricing models, ten messages or twenty UX flows, the “user interface” may no longer be a screen at all. Testing might be to find prompt(s) to AI Agent(s) deliver needed outcomes.
Ok, but this person didn't even seem to be doing enough paying to the market of one version already?
And while this claim about parallel development being a huge unlock is the most interesting thing, it also sounds a bit glib. Getting your foot in the door is the hardest thing early on, now you're trying to run six versions of your company at once? Each time you get a foot in the door sales-wise, are you trying to make them use all 6 versions, or are you only gonna get feedback on 1? Would you want to pay money to be a beta tester of 6 different products simultaneously, with reason to believe that 5 of them will probably evaporate over night soon?
Seriously, where do ideas like this come from? An "ultracapitalist" country that has about as much redistributive social spending as other developed economies[0]? A "dystopia" that millions of people from all over the world clamor to get into every year?
[0]: https://www.piie.com/publications/policy-briefs/2016/true-le...
The man is an economist, not a crony operating at the federal level (one does not imply the other, and I know nothing of the man's background).
I don't agree with the parent; I think capitalism is doing a lot of great things for us and will continue to, even with AI. But man I'm tired of these hot takes from people with limited practical experience.
I hear you there. There's a reason I limit what I say online. I know very little about very little.
One year ago models could barely write a working function.
One year ago, the models were only slightly less competent than today. There were models writing entire apps 3 years ago. Competent function writing is basically a given on all models since GPT3.
Much of the progress in the past year has been around the harnesses, MCPs, and skills. The models themselves are not getting better exponentially, if anything the progress is slowing down significantly since the 2023-2024 releases.
It takes far less manual prompting to make it have consistent output, work well with other languages, etc. But if you watch the "thinking" logs it looks an awful lot like the "prompt engineering" you'd do by hand back then. And the output for tricky cases still sometimes goes sideways in obviously-naive-ways. The most telling thing in my experience is all the grepping, looping, refining - it's not "I loaded all twenty of these files into context and have such a perfect understanding of every line's place in the big picture that I can suggest a perfect-the-first-time maximally-elegant modification." It's targeted and tactical. Getting really good at its tactics for that stuff, though!
I can get more done now than a year ago because taking me out of the annoying part of that loop is very helpful.
But there's still a very curious gap that the tool that can quickly and easily recognize certain type of bugs if you ask them directly will also happily spit out those sorts of bugs while writing the code. "Making up fake functions" doesn't make it to the user much anymore, but "not going to be robust in production but technically satisfies the prompt" still does, despite it "knowing better" when you ask it about the code five seconds later.
That has not been my experience. This weekend I pointed Claude Code+Opus 4.6+effort=max at a PRD describing a Docusign-like software. The exact same document I gave to Claude Code+Opus 4.5+Ultrathink around 6 months ago.
The touch-ups I needed after it completed implementation was around a tenth that it took with 4.5. It is a pretty startling difference.
If a metric goes from 0 to 2 it doesn't mean it's on a long-lived exponential trajectory.
This is a false claim.
Claude Code was released over a year ago.
Models have improved a lot recently, but if you think 12 months ago they could barely write a working function you are mistaken.
Even if a lot of the improvements we see today are due to things outside the models themselves -- tools, harnesses, agents, skills, availability of compute, better understanding of how to use AI, etc. -- things are changing very quickly overall. It would be a mistake to just focus on one or two things, like models or benchmarks, and ignore everything else that is changing in the ecosystem.
We can’t say for sure yet which trajectory we are on.
90% of blog articles created in the last two years are probably dead on arrival
There's live-coding, so it's not totally a crazy idea.
Not sure how to make that a platform, as when i wrote i explicitly put everything directly into a book unedited, whereas for many people, the editing is probably at least half if not more of the time they spend writing.
Or we could just bring back Google Wave :-)
jimkleiber.com/project-35 if you’re curious.
> The bottleneck is no longer engineering. It’s moving up the stack to judgment, customer insight for desired outcomes and distribution.
He's not saying "add AI to your product" or "use AI or die" but more that AI has shifted institutional assumptions about tech stacks, defensibility and fundability. The bottleneck moved up the stack from engineering to judgment, insight and design.
Chris lost because he was heads down building while $20B in defense VC was flowing into his exact problem space and he didn't build the boat to capture that wave.
The big pinch of salt I throw in with advice like this though is that startup failure rate hasn't dramatically shifted despite two decades of lean startup methodology, accelerators, and an entire cottage industry of startup advice. It's never the fault of the framework, mind you.
edit: Tobi from Shopify has an insight that relates. His north star metric is user churn. sounds crazy on its face. he's known for that. But increasing churn means you've increased top line exposure to more would-be entrepreneurs. Not all of them will succeed, but shopifys mission is to create more entrepreneurs. Grow the pie. A focus on increasing conversion tends to have a narrowing effect.
If you have good product idea, the methodology to get there mostly affect profit marigins, not whether it will be success or total failure
What happens is that the original idea rarely matters at all. It is the people that implements the idea what matters.
The original idea is almost always terrible, but great people pivot or change the idea gradually while having contact with reality.
You’ve always needed to constantly learn and innovate to launch a successful business.
Startups are mostly all default dead. That's why they need VC money.
Of all the things that AI has changed, tech stacks aren't one of them. The bots will gladly write Typescript, Java, Python, Rust, what have you. They could not give less of a shit.
What is he getting at? How does the code and infra stack differ at all between a company that is using AI, vs one that is not?
Build vs. buy is an eternal question in enterprises. I remember many in-house data teams trying to build tools for "digital transformation" and cloud migration about 10 years ago. The challenge was, building those tools was more expensive than those enterprises could budget for (IT as cost center), so a startup like Snowflake would easily outcompete in-house solutions with their custom, cloud-based tech stack that was necessarily complex because it needed to serve the needs of thousands of customers.
If he's right, the build vs. buy equation has shifted more towards build, at least as far as enterprise software is concerned. IT is still a cost center, but in theory an internal team can now handle more requests for custom tools without looking to outside vendors. Essentially the cost of building in-house might be collapsing and therefore enterprise software startups will be serving fewer customers (who would all pay you more because if solving the problem was cheap they'd do it).
If you had to build a stack for dozens of customers paying huge amounts of money, how would that stack differ from the stack you'd build to serve thousands of customers? Certainly it wouldn't need to be as scalable! And that's probably what he's getting at. I think what you'd do instead, to capture those higher price point customers, is solve their problems more specifically, in a higher value manner.
Many companies already do this, investing far more in field engineers than they do in their tech stack, since customization is essential.
It is also not the same as, "If you want to be a profitable company...". For that you need to somehow make more money than you are spending.
I've had this idea of 'business as reducing entropy' floating around in my head for awhile. It's a neat way to think about the value a business offers to buyers; a washing machine manufacturer is selling reduced time to reduced entropy (clean cloths), spreadsheet software is selling reduced time to understanding (information from tabulated data), and so on.
From that perspective, a lot of AI-driven development is failing.
We're still in the phase of 'how do we get order out of semi-average chaos?' for LLMs. For ML we're largely past that point.
I've been using this framing as a means to guide me towards 'what is actually useful, what might someone actually buy'. I don't have my own business at this point, but its still fun to think about off and on.
You could generalize this to the purpose of life itself, probably.
I disagree with this.
On pricing, I get that agents and tokens can scale in a way that's unrelated to # of users. But for much SaaS software, AI remains helpful to a human and the human remains the receiver of value. Seat-based pricing is easy to understand and you can always layer in token/agent costs thresholds.
On features vs. outcomes, the latter is hard to define and measure in many industries. In marketing SaaS, which I know well, you can't often tell what outcome to expect. You have to try a lot of ideas and some will hit. No way a SaaS vendor can guarantee that.
Not sure I agree with this train of thought, but a SaaS CEO made this exact argument to me last week.
VC for conventional SaaS is dead.
My posit is this: engineering never was the bottleneck, or at least hasn't been for 10 years now. Frameworks and best practices are pretty well known at this point. AI is simply exposing this reality to engineers' faces.
Proof point - most publicly traded SaaS first businesses S&M equals their R&D spend, if not dwarfs it. You're going to see this even more lopsided going forward.
The author didn't spend more than, maybe, 30 seconds thinking this through? Information I could've gotten in 3 seconds by opening a screen and looking at a line item, I now have to extract by writing a paragraph to an AI agent (and cross my fingers that nothing I said was ambiguous or misunderstood). And that's supposedly an upgrade?
That said, if you believe universe exists, chances are not null that you are correct. But solipsism might actually be right.
In case of doubt, remember that your memory might be mere illusions.
This trap has killed many startups, well before AI.
Now that code is cheaper to write, hopefully it becomes less of a problem?
In either case, founders should never fall in love with their solutions.
It also fails to convey that he's actually only talking about startups that were created 2+ years ago, rather than the many AI startups founded in the last 2 years.
Hahahahahahahaha no you can't. The rise of LLMs has done little to nothing in this area because it's very much compute-limited. Digital-twins and other ML-based strategies predate ChatGPT by a long shot. There are definitely places in hardware design where LLMs and agentic workflows will help, but that's largely because the existing tooling is utter garbage, and now the industry has a fire under its ass to make things automatable so they can build their own agents.
Now with AI, it is likely going to be 98%.
After AI: 4900 of 5000 fail (98%), 100 succeed
Like this?
2021-2024: good time in US for EV startup
2025: terrible time in US for EV startup
2026 March/April: AWESOME time in world for EV startup
focus on fundamentals, not flakey ephemerals
2020: wise to have smart elite software engineers on your team
2021: ditto
2022: ditto
2023: ditto
2024: ditto (is this when ChatGPT launched? dont care. snore)
2025: ditto (what are YC/HN/VC hyping now? snore)
2026: ditto
2027+: ditto, likely
In that case, yes their startup is most certainly DOA.
>Most AI users out there are Q&A'ing it and they have no idea what agents, tool calling or context compaction are.
Again, talking about a tech CEO not a random "AI user."
- The AI jump happened in Q3-Q4 2025 with Opus 4.5 so it's been six months or so? Not long enough.
- Most developers out there use AI for their coding work, not for re-envisioning business models.
Going to VCs with a 2 year outdated deck with no AI functionality or plans or tool use is unimaginable.
As to my original point I'll go and say that 95/100 of business owners out there in the world don't have any idea about Claude Code, Codex or anything of that caliber. It's too early. By the time that group gets to it there'll be tools tailored to their needs, not a terminal-based coding agent messing up filesystem.
You keep returning to this which is a strawman of my argument. From the start I have been talking about technical startup founders, which is also the subject of the blog post.
I'm not assuming anything, that is my argument. I think you could do a better job countering it than pointing at the status quo of every business in the world. lol.
Me: 'cutting edge startups should use cutting edge tools'
You: 'every business in the world isn't using cutting edge tools!'
If your business is selling services at 40% margin that are entirely digitally based, then maybe you’ll need to cut some margin, sure.
