chased

we are now 40% to the top, load your bags and sell at 1TN.

@teortaxesTex he can just tell lies and make money? worked for trump

He can't outscale Google in compute. Can't outpoach Meta, can't out-culture Anthropic, can't out-publish GDM. The management culture is so retarded that everyone left. He hates "researchers" so there's no program. xAI is a product burden. Imagine is a turn-off What does he offer?

Honestly, I don't see how xAI makes it. Not a hot take, just genuinely confused. Elon is trying to reduce it to an overpromise-and-scale-hardware-problem, on multiple dimensions, because that's his only forte. And he's got no Gwynne Shotwell. The whole thing is just bizarre.

@tszzl @karpathy @soumitrashukla9 if by non technical people you mean lil baby, meek mill, baby keem etc then absolutely.

@karpathy @soumitrashukla9 non technical people are downloading something called openclaw and using it in their terminal?

Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

@_TarunKathuria This post would go so hard on blue sky

The probability that Elon actually knows the size of Claude models is less than 25%. The probability that he knows the size of Grok models is less than 5%

@LarryGold310597 @jukan05 Lol “Repeating the old playbook in a changed environment isn't sophistication — it's a failure to price in structural change.”

@jukan05 You need to use LLm's less in your writeups. Right now it's glaring.

Dan wrote a good piece, but I see things differently on a few important points. 1. Buyer resistance to smartphone memory price hikes is largely confined to legacy memory like DDR4 — not memory broadly. This distinction matters. It's true that some buyers have recently pushed back on price increases, but the target of that resistance has primarily been legacy memory like DDR4. The market has been behaving rather abnormally of late. At one point, legacy memory like DDR4 was actually trading above DDR5 in price — a genuinely strange phenomenon. That distortion has since moderated as DDR5 prices have risen, but the spike at the time was difficult to explain through normal demand alone. According to industry sources, a significant portion of the DDR4 price surge was attributable to Chinese stockpiling. That's what gave smartphone OEMs the room to respond — by downgrading specs on entry-level devices or trimming production volumes. In other words, there was flexibility on the legacy side. DDR5 is a different story. As I noted in an earlier post, smartphone and PC makers accepted substantial DDR5 price increases in Q1 of this year, and even into Q2. The implication is clear: DDR5 is not a negotiating target right now — it's closer to an essential input that buyers need to secure even at a premium. What's more, flagship products built around DDR5 aren't easy to spec down. At most, a spec freeze is on the table. I've also heard some buyers say they simply can't keep absorbing additional costs on legacy memory. What I failed to account for was that most market participants weren't drawing a clear enough distinction — that this dynamic was specific to legacy — and I didn't anticipate that gap in understanding. 2. Repeating a familiar playbook and adapting to structural change are two different things. Dan portrayed investors who reflexively sell on spot price declines as sophisticated. I'd disagree. Memory companies are no longer operating the way they used to. If anything, they're moving toward a model closer to what TSMC does — building out capacity after securing advance paymets and long-term demand visibility from key customers. Just a few days ago, Korean media reported that Samsung is in discussions to pursue prepayment-based agreements with the likes of Microsoft. And yet the market still treats memory companies as textbook cyclicals. Even as contract structures and demand visibility that resemble TSMC's begin to emerge, the Big 3 trade at valuations that remain absurdly cheap by comparison. Repeating the old playbook in a changed environment isn't sophistication — it's a failure to price in structural change. I'm well aware of the counterargument. LTAs existed in prior cycles, and many were torn up when downturns hit. But the memory companies know that better than anyone. There's no reason to think they don't want TSMC-like valuations. And precisely because of that, the incentive to avoid repeating the old pattern — overcapacity that destroys their own cycle — is far stronger than it's ever been. They're working considerably harder than most people realize to pursue disciplined expansion rather than reflexive overbuilding. It's disappointing to see so many participants still running the same old framework, watching spot prices tick on a screen rather than paying attention to how the industry itself is changing on the ground.

@jukan05 Reads like pure ChatGPT slop

@OldInternetFeel Rare Pete Davidson

@TVietor08 The present is we get to send troops to die for kharg island

Anyone know wtf this is this about?

@0xMerp Ever since the DoD switched to ChatGPT from Claude, Pete’s speeches have gotten way more cringe

Lmfao.

@robuhst @jimmyjerkit @Okami13_ lol

@StayingChased @jimmyjerkit @Okami13_ They have meaningless daily fetch quests you need to do every day so you can get exp and currency

Crimson Desert has a staggering amount of content for the 100%. • 76 unique bosses • 467 unique characters • 29 different mounts • 110 different factions • 401 creatures to discover • 355 crafting blueprints • 573 territories to explore • 150 different materials • 430 feats • 94 secret collectibles

@jimmyjerkit @Okami13_ The game started development as an MMORPG and they just switched it to single player later on so I imagine it’s mostly MMO slop

@Okami13_ You cannot tell me that more than half of that isn’t meaningless filler.

@tszzl @pmarca This is only true if you really think about it

@pmarca an entire book where the guy is introspecting

Marcus Aurelius, Meditations, X.16: “To stop talking about what the good man is like, and just be one.”

next: Jalisco cartel vs the Discombobulator

@josh_uglyasf It’s double-barrel Cheryl

Saw what now 🧐🧐🧐

@flavioAd My momma is crying, she got me crying over this deprecation. I've never been a personal fan, but my mom was practically raised by his sycophancy, in her dark and happy moments through life. R.I.P GPT-4o 🖤

GPT-4o is officially being retired on February 13.

@thebigmehtaphor @OpenAI My momma is crying, she got me crying over this deprecation. I've never been a personal fan, but my mom was practically raised by his sycophancy, in her dark and happy moments through life. R.I.P GPT-4o 🖤

unfathomably cold of @OpenAI to retire GPT-4o the day before Valentine's

@loopify Elon’s net worth ain’t 1T yet

that was the last exit pump gut telling me market top

@carlhancock @markgurman Even the dumbest ChatGPT answer is still a better one than Siri can scrounge up

@markgurman Apple doesn't think consumers don't want chatbots that provide incorrect answers or help people off themselves. That is their conundrum. Along with wanting it to work locally and be privacy focused. Despite what you and others seem to think.

OpenAI says ChatGPT has 800 million weekly users. It continues to be mind-boggling that Apple is sticking to its line that consumers don’t want chatbots. In less than three years, it has become one of the most used products in world history.

@Tawadros15 Since the holy sea is its own country with its own bank they could have more money in the bank than we know about right? Hypothetical I know but I don’t know how it all works

@StayingChased No doubt but you can't and it wont ever be sold.

People hate the Catholic Church for supposedly hoarding wealth and being obscenely rich. The Catholic Church has around 15 billion dollars in reported assets for a religion of 1.4 billion people. That's $10.71 per Catholic. You hear people go on and on daily that this is unacceptable wealth. The Church of Jesus Christ of Latter Day Saints, aka Mormons, has 265 billion dollars in reported assets for a church of 17 million people. That's $15,588.24 per Mormon. You rarely, if ever, hear anyone talk about this. There are 5,500 Catholic hospitals, 221,000 Catholic k-12 schools, 9,000 Catholic orphanages, 1,300 Catholic colleges and universities, over 160 charitable organizations and estimated 500,000 local charity offices in the world. There is only 1 hospital Intermountain Health in Utah founded by Mormons that no longer has anything to do with the LDS, 15 Mormon k-12 schools, 4 Mormon universities, 0 Mormon orphanages, Mormon charities is under Deseret Industries with less than 10,000 local centers.

@ai_for_success if sam is posting stuff like this its probably because some of his sweet little lies hes been telling internally are close to causing a rift between jakub and szymon

Sam Altman : There are two people I'd like to mention that OpenAI would not be OpenAI without: Jakub Pachocki and Szymon Sidor. Mark Zuckerberg rigth now.