hzapperz

@itsjoaki This is the most retarded thing I’ve read in a while.

most vibe coders are just typing prompts into someone else's memory real leverage is knowing when to stop using an LLM and build it yourself. 🦞

@ondrejslunecko Game theory wise, if you have a 1.01% chance of a 100x you should take that bet.

Jeff Bezos - “If you have a 10% chance of a 100x return, you should take that bet every time."

Everything is downstream from distribution

@Midnight_Captl Yeah, you have to teach people how to use it, which is kinda surprising

A friend at Apple told me that over the past couple weeks, their team got access to Claude with a $300/day token budget. This is global sourcing on the business side, not engineering. I’m also hearing that when directors ask for backfill, senior leadership is asking what the team’s AI usage looks like. If token usage is low or nonexistent, the answer is increasingly: go figure out how to get more leverage out of AI first. Wild.

@BacLeodiv Yeah, requires a lot more steering than before

Opus 4.6 feels dumber lately just me, or anyone else feel the same?

@kalyan_wtf A better way of saying this would be: frontend is more tedious than backend

FRONTEND IS HARDER THAN BACKEND

@jacobrodri_ Mutual combat is allowed in the great state of Texas.

i still think a tinder but for fights could easily be a $100k/month app

@radmadvlad q2 kinda useless

single 3090 and dual channel ddr5 guess the model.

AGI is 20-200/month and you’re complaining?

@ns123abc $300 is more than enough in a day, if you’re using more than this, there’s a serious skill issue.

>Friend at Apple told me they’re being given $300 worth of token budget per day lol LMAO

@fitcapdefi The dosage appears to be 5 days on 2 days off. That’s what I’ve been told to do for Klow. I dose every 5 days or so.

also started 1mg of ghk-cu daily last week seemed to tolerate it well gonna bump that up to 2mg a day

@quinnslcm @skylarromines Sure, the people who watched it liked the video, but you said not enough people watched it so…

@skylarromines That’s not what 99% of people said who watched, they said the opposite

Although I think this sounds great, I did this for 100+ days straight and never broke over 300 likes. Got a lot of good feedback from the people that did see though Maybe I was too early?

@bubbleboi There are only a handful of ai chip makers, Google, Nvidia, and Amd. x ai coming soon. My approach has always been not to sell ai like a wrapper, but rather how do I make money using ai. I’ve increased my productivity therefore my labor is worth more than others.

I’m a big believer in not selling chips but selling the services those chips create. It’s the reason why cloud services is a larger market than CPU’s and why I think every AI chip maker will pivot to just serving and hosting models as a service rather than selling hardware which is a terrible business.

@tolstoybb You’re a year late.

One of my more crank beliefs is that Hitler revisionism is going to go mega mainstream in my lifetime.

@maxkolysh Thats not really what AWS and GCP do… Anthropic is not going to host websites them selves, they are going to use cloud providers to do so.

@zappstudios wdym? if you created an app in anthropic and it was one click to host with them, why wouldn’t you?

so does anthropic just eat AWS and GCP?

@DmytroKrasun You’re putting Cloudflare and HubSpot in the same bucket… Ngmi

It is not a software meltdown, it is a mental meltdown in the markets, also known as AI psychosis. Take these companies, check what they do, their moats, their customers, their revenue, and what their CEOs are doing now. Cloudflare, HubSpot, and Snowflake. What will likely happen to them and to businesses like that? I might be wrong, but markets seem to price narratives now, not fundamentals. It seems it is better to just quietly bet on my understanding of reality.

@bymarcoperez Bro made a newsletter into a mobile app.

My first iOS app has been live for 5 weeks. Here are the stats so far ⚡️ - 1000+ downloads - $194 MRR - $366 All-time Revenue - 30 five star reviews Initial launch hype has died down, but progress is progress!

@GaryMarcus CC is AGI. Too much goal post shifting has happened and there’s no real consensus on what AGI actually means. If you think otherwise, you’re probably not using Claude correctly.

Claude Code is not AGI, but it is the single biggest advance in AI since the LLM. But the thing is, Claude Code is NOT a pure LLM. And it’s not pure deep learning. Not even close. And that changes everything. The source code leak proves it. Tucked away at its center is a 3,167 line kernel called print.ts. print.ts is a pattern matching. And pattern matching is supposed to be the *strength* of LLMs. But Anthropic figured out that if you really need to get your patterns right, you can’t trust a pure LLM. They are too probabilistic. And too erratic. Instead, the way Anthropic built that kernel is straight out of classical symbolic AI. For example, it is in large part a big IF-THEN conditional, with 486 branch points and 12 levels of nesting — all inside a deterministic, symbolic loop that the real godfathers of AI, people like John McCarthy and Marvin Minsky and Herb Simon, would have instantly recognized.* Putting things differently, Anthropic, when push came to shove, went exactly where I long said the field needed to go (and where @geoffreyhinton said we didn’t need to go): to Neurosymbolic AI. That’s right, the biggest advance since the LLM was neurosymbolic. AlphaFold, AlphaEvolve, AlphaProof, and AlphaGeometry are all neurosymbolic, too; so is Code Interpreter; when you are calling code, you are asking symbolic AI do an important part of the work. Claude Code isn’t better because of scaling. It’s better because Anthropic accepted the importance of using classical AI techniques alongside neural networks — precisely marriage I have long advocated. It’s *massive* vindication for me (go see my 2019 debate with Bengio for context, or to my 2001 book, The Algebraic Mind), but it still ain’t perfect, or even close. What we really need to do to get trustworthy AI rather than the current unpredictable “jagged” mess, is to go in the knowledge-, reasoning-, and world-model driven direction I laid out in 2020, in an article called the Next Decade in AI, in which neurosymbolic AI is just the *starting point* in a longer journey.* Read that article if you want to know what else we need to do next. The first part has already come to pass. In time, other three will, too. Meanwhile, the implications for the allocation of capital are pretty massive: smartly adding in bits of symbolic AI can do a lot more than scaling alone, and even Anthropic as now discovered (though they won’t say) scaling is no longer the essence of innovation. The paradigm has changed. — *Claude Code is plainly neurosymbolic but the code part is a mess; as Ernie Davis and I argued in Rebooting AI in 2019, we also need major advances in software engineering. But that’s a story for another day.

@DrCamRx Peptides stimulate your bodies natural functions. This post is like debating if water is wet.

@heygeorgekal @claudeai Skill issue.

Been a @claudeai fanboy for a long time but the usage limits even on claude max are getting ridiculous Paying for both 100 claude and 100 codex this month. The difference is night and day. - Codex ($100): running parallel agents, barely touching the limits - Claude ($100): 3-4 deep prompts and i'm locked out I get it, they're probably focused on enterprise, but if you're charging $100/mo and people hit the wall after 30 minutes, they will churn. The product is incredible. The packaging is not.