Tweet Disematkan
You have to have a fanatical conviction in your own abilities bc why wouldn’t you
English
∿spencer.
12.3K posts
∿spencer.
@_ontologic
vice president of @conceptcountry // cohost of @ai_rebels // programmer, poet, poster
Bergabung Haziran 2021
1.1K Mengikuti2.9K Pengikut
The point of almost all of his books is not “god is dead” it’s “humanity can take you places where god can’t save you”
English
The Road is essentially a founding myth about the rebirth of civilization through a messianic figure in the wake of catastrophe. I don’t know how you could possibly believe him a materialist
Dwayne Barrick@SitLibertas
@christopherrufo I believe McCarthy was largely a misanthrope and an extreme materialist. He was a first-rate writer in terms of command of the language. But he did not believe it was worth noting the inner lives of his characters. He was Gnostic but without the desire for transcendence.
English
You should be thinking hard about how you get AI written code to fail as fast as possible
English
@graphtheory @shinboson It’s why he and I get along tbh. I have the same condition
English
thcv is great for writing but then the next morning I talk like sephiroth and no matter how hard I try, I cannot stop. my inner child is an edgy thirteen-year-old boy with very long bangs wearing a hoodie with those holes in the wrists
English
@shinboson They’re real mad about this one
English
high-openness right-leaning is just called being extremely tribal. you're functionally an orientalist for the fine china and exotic spices of cultures that produce works of art that your spiritually gelded one cannot.
alco ⊢ ꙮ@qualiascript
being right-leaning and high openness is so funny. "this is one of my favorite musicians, i disagree with everything they stand for, highly recommend"
English
It’s very cool. Been playing with it for the last day and I will never again use ollama or lm studio
∿@somewheresy
Actual Computer is inference mesh software that allows you to use a single terminal command to automatically set up clusters from most devices, with a passthrough for OpenResponses compatible APIs and remotely accessing your local compute cluster
English
I think ppl are overfocusing on model behavior and not zooming out to consider broader engineering principles and how they should be enforced on stochastic actors
English
Types are not the guardrails, types enable the guardrails
Michael Hart@hichaelmart
Kinda fascinated how ppl keep trotting out "typed languages are better for codegen" even though we have evidence against it. It *feels* intuitively correct, right? They give better guardrails! But maybe types aren't all that amazing as guardrails for LLMs? dev.to/mame/which-pro…
English
Urgh too much knowledge in my head think I might kill my self
- Nick Bostrom on infohazards
English
I’m an enormous AI optimist, but some ppl on here are insane and I question their relationship to information as a whole
English
@DeepDishEnjoyer It’s over isn’t it?
English
hahahahahahahahahahahahahahahaha
OSINTtechnical@Osinttechnical
Trump floats the idea of the U.S. abandoning the Strait of Hormuz.
Filipino
@word_and_number @somewheresy @sevensixfive There are people who still believe in establishing this and they are rife on this site.
English
QME
The name "Duncan Idaho" just gets funnier and funnier the more you encounter it
English
The cashier at Taco Bell was rude but she can’t faze me. I have my Evening Energy Rise Relax playlist on and am serene as a baby lamb nestled in its mother’s wool
English
∿spencer. me-retweet
∿spencer. me-retweet
you’re pitching garry tan
“so what do you guys do”
you start explaining
he’s furiously typing . two keyboards. one hand on each. you’ve never seen this before
“who are your top customers”
you explain. he types. his apple watch is a strobe light of notifications
“who’s your competition and why should i invest”
you explain that there’s no competition and you are the best and only product in the space
“false!” garry jumps out of his seat
“i am the competition!”
you are speechless
“in this meeting, i vibe coded your entire company. and my gstack has already closed your top customers.”
you check your phone. your stripe graph shows 100% churn
“and look at this”
garry shows you his imessage. there’s a text from 35 seconds ago. your top enterprise prospect that you’re trying to close? garry’s AI is trading baking recipes with the CEO’s mom
“thank you for playing!”
you have no moat. you are not admitted to the YC spring 26 batch.
“paula”@paularambles
garry in his office in his lobster outfit "okay claude... rewrite this but in rust... no wait... rewrite it like paul graham would" "garry you have a yc interview starting in 5 minutes" "one second. claude just one-shotted a distributed system" "garry they are in the zoom" "can they describe their startup to claude instead" "garry you are the interviewer" "hold on. claude says their idea has a better moat if we pivot them" "they haven't even pitched yet" "claude already knows" "garry this is yc" "...what's yc again"
English
Basically with AI code you gotta find ways to make it fail as fast as possible, in as many ways as possible. There will always be edge cases neither you nor your code agent consider. If you want velocity you need verification and observability
∿spencer.@_ontologic
Some things I’m doing to deal with this: 1.Built an MCP that automatically fuzzes and analyzes generated code, which helps catch edge cases traditional tests miss. 2.Putting more weight on integration and e2e testing, since a lot of the failure modes show up at boundaries.
English
I wonder if there’s a podcast where people have been talking about this exact idea for years @ai_rebels
Tyler@rezoundous
The real skill in the AI era might be deciding what not to build.
English
Some things I’m doing to deal with this:
1.Built an MCP that automatically fuzzes and analyzes generated code, which helps catch edge cases traditional tests miss.
2.Putting more weight on integration and e2e testing, since a lot of the failure modes show up at boundaries.
William Huster@whusterj
I am developing a formal theorem I call the Verification Complexity Barrier. In a nutshell, if a program has some components `n` that have connectivity factor of `k > 0` , then verification complexity increases superlinearly for each new component. Therefore the time required to fully verify the system always exceeds time to generate components. After a while, because it's superlinear, the verification complexity takes off and becomes impossible to keep up with in some finite amount of time. This was true before AI, but is much starker now as code generation time trends towards zero. You hit the barrier sooner. We all have finite capacity - even AI agents - so there will always be a certain number of components `n` where the wall is hit. You must spend more and more effort on verification for each new component in the system. The best thing you can do is spend time changing the "topology" of the problem - change the software architecture - so that the exponent of verification complexity is lowered ([1] Lehman) and the curve is flattened. You can bundle components into modules, you can add automated tests, you can use formal proofs, you can use type systems. These things push the barrier to the right. They buy you more components and a more complex system. But the theorem suggests you can only ever defer the barrier, never completely eliminate it. AI Agents can burn tokens all day long generating software components and tests for those components. They can find bugs and fix them, but they cannot prove the absence of bugs (per [2] Dijkstra, [3] Rice, [4] Smith, and others). And "Bug" needs to be defined against someone's spec for what "working" and "not buggy" looks like. The more you build with Agents, the heavier the verification burden becomes. This may sound like the same trite observation others have made here on X: "Our bottleneck is no longer writing code, but reviewing code" [5] and "I am the bottleneck now." [6] True, but I don't think anyone has captured the magnitude of the problem. The math is: at a certain component count `n`, it is literally impossible for you, your team, and your agents to completely verify a system. The bottleneck goes to zero, and nothing gets through. So what software companies do in the real world is release incompletely verified software and massively scale up. This shifts the burden of verification onto their customers, because "given enough eyeballs, all bugs are shallow" ([7] Raymond). If you can get enough eyeballs, this is a very cost-effective way to shift the barrier to the right by massively increasing your team's capacity. You walk the tightrope of doing enough internal verification before release so you don't lose customers, while tolerating a certain amount of escaped bugs, which - if those bugs matter at all - your customers will find for you. Meanwhile, massively scaling up just accepts the growing cost of complexity. You can push `n*` from 15 to 30 by quadrupling your capacity. To get to 60 you need to quadruple again, and then again to get to 120. Your cost curve is superlinear to get linear gains in system size. At a big enough scale, you amortize the cost across your customer base and the economics work. Contrast that with a sufficiently complex vibecoded app built for a small audience - high complexity costs can't be amortized at small scale. I expect to see many people and companies try and fail at vibecloning complex SaaS in the near term. Complexity cost economics only scale with audience size (I will share another model for this). I do think SaaS prices will be corrected downwards to account for savings in code generation, but I predict that once the irrational exuberance for vibing fades, we'll see that it still makes sense to buy rather than self-build complex SaaS. A broader implication is that AI Agents will never be able to self-verify. Humans, too, will never be able to fully verify their behavior, because LLMs are by design of maximal complexity. Did you see the size of those error bars in the latest METR results? [8] The longer the horizon on a task, the more spread in AI agent outcomes. This is the Barrier in action. Spread is a feature of GenAI, but in practice it means heaps more output to review and verify. The Complexity Barrier shows you literally won't have time to review it all. At the inflection point of verification complexity, you have to fall back on vibes. The implication for fast-takeoff AGI is even scarier: if AI does reach a point of recursive self-improvement, this theorem suggests it will be structurally impossible to know that behavior is aligned, because you won't be able to completely verify. Drift is bad enough in vibe coding. Runaway AI will drift massively and there's no way of knowing where it will end up. All that's to say, verification should be the focal point of AI Engineering for the foreseeable future and maybe forever. That is: how do you capture what you want to do, refine that into specifics, and then follow up with automated tests, assertions, evals, and customer feedback to progressively harden your software? The verification problem is acute now because of how cheap software generation is. Because of the superlinear nature of software verification complexity, companies that push hard on the barrier and successfully shift it right will have a built-in moat versus those who fail to put in the verification work. Detailed blog post and interactive model incoming. Links: [1] en.wikipedia.org/wiki/Lehman%27… [2] cs.utexas.edu/~EWD/transcrip… [3] en.wikipedia.org/wiki/Rice%27s_… [4] cse.buffalo.edu/~rapaport/Pape… [5] x.com/shl/status/194… [6] x.com/thorstenball/s… [7] en.wikipedia.org/wiki/Linus%27s… [8] metr.org/blog/2025-03-1…
English
@SashaGusevPosts It’s easier to make sense of when you start from the understanding that all these eugenics freaks have a breeding kink and ultimately are just trying to sane-wash their political belief that the state should control who has sex with who
English
It's amazing how much hereditarian discourse is just them not knowing math, thus making discourse impossible.
eigenrobot@eigenrobot
@SashaGusevPosts thats not how any of this works
English