@Dr_Gingerballs Avatar @Dr_Gingerballs Dr_Gingerballs

Dr_Gingerballs posts on X about ai, this is, open ai, money the most. They currently have [------] followers and [---] posts still getting attention that total [------] engagements in the last [--] hours.

Engagements: [------] #

Engagements Line Chart

Mentions: [--] #

Mentions Line Chart

Followers: [------] #

Followers Line Chart

CreatorRank: [-------] #

CreatorRank Line Chart

Social Influence

Social category influence technology brands 12.71% finance 12.15% stocks 3.87% cryptocurrencies 2.76% celebrities 0.55%

Social topic influence ai #2803, this is 6.08%, open ai #1327, money 3.31%, if you #4893, we are 3.31%, in the 3.31%, vibe coding #1049, future 2.76%, anthropic 2.76%

Top accounts mentioned or mentioned by @ser_kingslayer_ @andyxandersen @todistetta @georgee33179 @nicholasdclay @ewinsberg @jadecole2112 @meanwhileinca @goldmarkovchain @sszone104322 @tmtlongshort @edzitron @yieldsearcher @vujademgmt @ednewtonrex @enghumanista @jasonfurman @siriuslybs @groby @chasebrowe32432

Top assets mentioned Alphabet Inc Class A (GOOGL) Microsoft Corp. (MSFT) Spotify Technology (SPOT) Bitcoin (BTC)

Top Social Posts

Top posts by engagements in the last [--] hours

"This one was a lot of fun to think about and to write and I hope it provides clarity into why LLMs are numbskulls: They are digital Tommy Boys. My next piece is going to go into art and creativity and why current AI models by definition cannot be artists and cannot create art. https://t.co/iaFmpiwh0Q https://t.co/iaFmpiwh0Q"
X Link 2026-02-11T01:56Z 24.1K followers, [----] engagements

"@jasonfurman Jon Stewart is one of the most thoughtful and nuanced interviewers in geopolitics I've listened to over the past [--] years. If he made an economist lose it it's probably because he drilled down on how intellectually bankrupt the field of economics is"
X Link 2026-02-11T02:04Z 24.2K followers, 33.1K engagements

"The AI benchmark cycle: [--]. Create new benchmark that all models fail. [--]. As details of the test slowly leak out train models on large sets of synthetic data. [--]. As scores improve prune the model and optimize it to perform well on that test at lower cost until your model just beats the other models. [--]. Wait for other model beats your score continue pruning and optimizing so you can beat that one on your next model release. [--]. Repeat until the test has been saturated. [--]. Create new benchmark that all models fail. Sundar buried the real story in the cost data. Gemini [--] Deep Think went from 45.1%"
X Link 2026-02-12T23:35Z 24.2K followers, 13K engagements

"@LevyAntoine Apple made income from their product. The AI labs lose more money than any company in history. Hope this helps"
X Link 2026-02-13T04:44Z 24.1K followers, 14.1K engagements

"Look at PPI numbers. A lot of companies are not passing on price increases for fear of losing market share. Small businesses are idiosyncratic like that. I wanted someone to come do some work on my pool that had good reputation back during covid. He said he couldn't take my job because he was too busy and couldn't find labor to support the growth. Of course he could have charged me double and I would have said yes and he could have passed that increase to wages to entice more workers to join. Last two nights I have gone to dinner and literally didn't care what the bill was. They could have"
X Link 2026-02-14T02:20Z 24.1K followers, [----] engagements

"Nobody thought heavier than air flight was impossible. People were studying lift and drag on airfoils a hundred years before Kitty Hawk. They thought it was difficult but possible. The challenges were technical not physical. Same with Edison and lightbulb. @theojaffee "AI cannot in principle make novel discoveries" is the new "heavier than air flight is impossible." the takes that age the worst are always the ones stated with the most confidence @theojaffee "AI cannot in principle make novel discoveries" is the new "heavier than air flight is impossible." the takes that age the worst are"
X Link 2026-02-14T14:40Z 24.2K followers, [----] engagements

"Because if they did that then it would be immediately clear how much copyrighted material they stole and would get sued into oblivion. Wait this sounds incredible useful Can we just have a model with [--] entropy [--] hallucinations that just acts like a retrieval database over its training dataset Also sounds like a great way to solve the traceability problem. Why don't the AI labs just make something like that https://t.co/n37BIqqQPO Wait this sounds incredible useful Can we just have a model with [--] entropy [--] hallucinations that just acts like a retrieval database over its training dataset Also"
X Link 2026-01-29T01:20Z 24.2K followers, [----] engagements

"Here's a great example. And I actually think both of the economists are overstating any evidence. They are both astrologists. But one of the first claims out of Burman was that higher fed funds rates result in lower inflation rates and lower unemployment rates. This cannot be concluded from the data. It simply can't. It's astrology. If anything if you ignore the last cycle higher rates are positively correlated with higher inflation. But what the last cycle shows us is that the correlation really isn't that strong and is likely driven by other factors that aren't controlled for by the fed."
X Link 2026-02-11T02:25Z 24.2K followers, [----] engagements

"This is the big clue to me. Looks like they heated the furnace up to initiate nucleation then turned it off and let it cool down while the flakes grew. Again not really sure what AI even did here"
X Link 2026-02-12T22:35Z 24.2K followers, [----] engagements

"This is what I mean when I say there is an unintelligent solution to every intelligence test. These benchmarks 😞 This is how you break them: [--]. Ask your most capable model to read a public task A and generate a similar task and then solve it. Given that the solution is verifiable you get a collection of task-solutions similar to public task A but not copying it. [--]. These benchmarks 😞 This is how you break them: [--]. Ask your most capable model to read a public task A and generate a similar task and then solve it. Given that the solution is verifiable you get a collection of task-solutions"
X Link 2026-02-12T22:52Z 24.2K followers, [----] engagements

"Funny this morning I encountered a wild UI bug that I had never seen before and wondered if they picked up vibecoding. I just love how software keeps disintegrating in the age of AI. Spotify says its best developers havent written a line of code since December thanks to AI https://t.co/6hafAJOeJv Spotify says its best developers havent written a line of code since December thanks to AI https://t.co/6hafAJOeJv"
X Link 2026-02-12T23:58Z 24.2K followers, [----] engagements

"I'm going to start a toe ring company and tell my potential investors that my TAM is 83B toes and 166B digits if we expand into fingers. Look at Anthropics latest raise. Look at the size and the names participating; funds of every kind. We should think rationally about this. Is it a bubble Perhaps. But Im choosing to believe theres something deeper at play here. The question at the end of the day is whether https://t.co/WHtmUqpiEM Look at Anthropics latest raise. Look at the size and the names participating; funds of every kind. We should think rationally about this. Is it a bubble Perhaps."
X Link 2026-02-13T04:35Z 24.2K followers, [----] engagements

"If we get enough people at OpenAI to see this post they will have to stay all weekend doing human assisted training to add this to the black box magic machine"
X Link 2026-02-13T05:03Z 24.2K followers, 23.1K engagements

"This is equivalent to creating an AI that can play chess. It didn't discover anything. The humans did that. It simplified and formalized the pattern the humans discovered. It's an important distinction because the humans did all of the thought experiments which is the hard part for rational beings. The easy part for computers is linking that physical argument to a string of operations to provide a larger sample framework. ChatGPT literally went from not being able to tell how many "Rs" are in strawberry to discovering new scientific breakthroughs The first-ever physics paper where AI acted as"
X Link 2026-02-14T05:04Z 24.1K followers, 37.4K engagements

"@TheStalwart Well the problem is you need to convince yourself that it is at least as accurate if not more accurate than alternatives and that work alone destroys the vibecoding work"
X Link 2026-02-14T05:18Z 24.2K followers, [---] engagements

"Hoo boy now we are getting into "this violates the 2nd law of thermodynamics" territory on the rationales for AI slop. A possible future: For the last [--] years devs have made the sensible assumption that every time a human touches a codebase under time pressure the codebase gets less maintainable. This is called software entropy. It's only possible to escape by continual monitoring and A possible future: For the last [--] years devs have made the sensible assumption that every time a human touches a codebase under time pressure the codebase gets less maintainable. This is called software"
X Link 2026-02-14T05:25Z 24.1K followers, [----] engagements

"@MrEwanMorrison The hole in the head is included for free with the implant"
X Link 2026-02-15T16:23Z 24.1K followers, [---] engagements

"@erikbryn I love economists. They are all theories in search of a reality"
X Link 2026-02-15T17:11Z 24.1K followers, [----] engagements

"@_everythingism I genuinely cant tell if its satire"
X Link 2026-02-15T17:18Z 24.1K followers, [----] engagements

"@HealthRanger Have you considered that you are a dipshit First question: did you read the paper. Second question: do you understand what the contribution from the AI was"
X Link 2026-02-16T00:46Z 24.1K followers, [----] engagements

"AI psychosis comes for us all"
X Link 2026-02-16T07:06Z 24.1K followers, 38.3K engagements

"@AIandDesign @HealthRanger Youre an AI thought leader. If AI doesnt work out you are sunk. Who is on the copium Snake oil salesmen disgust me"
X Link 2026-02-16T07:44Z 24.2K followers, [---] engagements

"I wanted to put all of my criticisms of AI as a potential replacement for humans into one place so people can understand how I evaluate new information about this space and hopefully be able to http://x.com/i/article/2009082694072733696 http://x.com/i/article/2009082694072733696"
X Link 2026-01-26T22:44Z 24.1K followers, 63.3K engagements

"Things are looking more and more grim for $BTC ETF holders. While holdings have been roughly flat at around $53B since June '25 price has dropped around 30%. Even Saylor buying ham-handedly cannot stop the drain. The drain largely comes from mining which has mined about 60% of all ETF deposits since inception. If investors ever go to collect their $53B they will find only $23B remaining. The break even price on ETF holdings is around $81K per coin. Saylor break even is around $76k. As far as an investment goes the returns have been a flop. If $BTC drops below $80k I expect investors will"
X Link 2026-01-30T15:26Z 24.1K followers, [----] engagements

"They somehow broke excel. A program that has needed no real innovations in [--] years. We are talking bugs that make it kind of unusable. Whatever thy are doing is a mess. 🦔 Microsoft is pulling engineers off new features to stabilize Windows [--] after months of patch failures. January brought emergency fixes for systems that couldn't shut down OneDrive and Dropbox freezing and machines stuck on black screens at boot. Windows chief Pavan Davuluri https://t.co/UCa41r05nf 🦔 Microsoft is pulling engineers off new features to stabilize Windows [--] after months of patch failures. January brought"
X Link 2026-02-01T15:00Z 24.1K followers, 348.5K engagements

"@RealJimChanos @RandolphDuke7 [--] years I have a space elevator to sell him"
X Link 2026-02-03T23:14Z 24.1K followers, [---] engagements

"Anthropic also took a big swipe at itself because the best use case it can come up with to market is a human asking how to speak to a parent. Groundbreaking stuff we all have that problem and cant come up with things like listen more. Anthropic just took a big swipe at OpenAI's decision to put ads in ChatGPT. Anthropic is airing ads mocking ChatGPT ads during the Super Bowl and they're hilarious 😅 Anthropic is also committing to no ads in Claude https://t.co/LR1v4xz9ds https://t.co/PXoaZtmCWA Anthropic just took a big swipe at OpenAI's decision to put ads in ChatGPT. Anthropic is airing ads"
X Link 2026-02-04T14:28Z 23.9K followers, [----] engagements

"@Redshiftty Perhaps the ad is a Freudian window into the souls of the people making Claude"
X Link 2026-02-04T14:45Z 23.9K followers, [--] engagements

"Folks my best guess estimate just from skimming top [--] BTC miners is they are on the hook for well over $30-40B in future AI compute capacity. If they cant get and pay their loans with AI revenue they have to sell BTC. BTC is AI now. I dont think its going well"
X Link 2026-02-04T19:39Z 24.1K followers, 20.9K engagements

"Is this good"
X Link 2026-02-04T21:13Z 24.1K followers, 15.6K engagements

"How dare Dario lie about my product Its my job to lie about my product First the good part of the Anthropic ads: they are funny and I laughed. But I wonder why Anthropic would go for something so clearly dishonest. Our most important principle for ads says that we wont do exactly this; we would obviously never run ads in the way Anthropic First the good part of the Anthropic ads: they are funny and I laughed. But I wonder why Anthropic would go for something so clearly dishonest. Our most important principle for ads says that we wont do exactly this; we would obviously never run ads in the"
X Link 2026-02-04T21:42Z 23.9K followers, [----] engagements

"They spent $20k to copy an open source repo. This tech is just so cool. Anthropic had [--] AI agents build a C compiler from scratch. 100k lines compiles the Linux kernel $20k [--] weeks. To put that in perspective GCC took thousands of engineers over [--] years to build. (Granted from [----] - however) One researcher and [--] AI agents just built a https://t.co/pk0PxsdTKi Anthropic had [--] AI agents build a C compiler from scratch. 100k lines compiles the Linux kernel $20k [--] weeks. To put that in perspective GCC took thousands of engineers over [--] years to build. (Granted from [----] - however) One"
X Link 2026-02-06T01:58Z 24.1K followers, 167.4K engagements

"Guy who has no domain expertise says vapid and provocative thing about said domain to funnel traffic to his newsletter. EVERY. SINGLE. TIME. for me the odds that AI is a bubble declined significantly in the last [--] weeks and the odds that were actually quite under-built for the necessary levels of inference/usage went significantly up in that period basically I think AI is going to become the home screen of a for me the odds that AI is a bubble declined significantly in the last [--] weeks and the odds that were actually quite under-built for the necessary levels of inference/usage went"
X Link 2026-02-06T02:17Z 24.1K followers, 25.5K engagements

"Im not convinced that alphafold has even done much meaningful predictions outside of the training data. Its really brilliant. Use the entirety of human knowledge in some domain to train a model so there is no data left to actually test the model. I wish I thought of that grift. https://twitter.com/i/web/status/2019654341397803286 https://twitter.com/i/web/status/2019654341397803286"
X Link 2026-02-06T06:08Z 24.1K followers, [----] engagements

"@macrocephalopod Not to be that guy but are you guys in quant land inverting matrices by hand Last time I did that was [--] years ago computers do it for you"
X Link 2026-02-06T06:14Z 23.9K followers, [----] engagements

"It just occurred to me that this will likely be the first case of model collapse where both the AI and the human operator models collapse in unison. Software development is undergoing a renaissance in front of our eyes. If you haven't used the tools recently you likely are underestimating what you're missing. Since December there's been a step function improvement in what tools like Codex can do. Some great engineers at Software development is undergoing a renaissance in front of our eyes. If you haven't used the tools recently you likely are underestimating what you're missing. Since"
X Link 2026-02-06T13:05Z 24.1K followers, 13.7K engagements

"Coming from a company that published a method to create non-consensual deep-fake Santa doll porn. I'm going to guess the SaaS tool they replaced probably just could have been deleted. But we will never know because the claim is so vague it cannot be criticized. Just like every other claim on this site. https://youtu.be/4tOzX0ELruksi=iCmiVp6HzLcUp5sV I work in a 1600-person company where we did replace an expensive SaaS product with an internally created piece of software built by a senior engineer using AI. So this is happening. In our case the software wasnt just equivalent but better for"
X Link 2026-02-06T16:07Z 24.1K followers, [----] engagements

"@bet_alexander7 Already done back in the crypto hype days. Look up SingularityNET"
X Link 2026-02-06T16:12Z 24.1K followers, [---] engagements

"@ryxcommar They stuffed shit in our postal boxes then stuffed shit into our email boxes then stuffed shit into our voicemail boxes then stuffed shit into our text message boxes now stuffing shit into the entire internet. Its all so anti-human. These people hate us"
X Link 2026-02-06T18:55Z 24.1K followers, [----] engagements

"Vibe coding is going well. 🦔 Microsoft's January Patch Tuesday update is now being blamed for slowing down Nvidia graphics cards. Users are reporting 15-20 fps performance losses flickering shadow artifacts and frame generation errors. Nvidia has confirmed they're investigating and advised users to https://t.co/iVC3EZxPNW 🦔 Microsoft's January Patch Tuesday update is now being blamed for slowing down Nvidia graphics cards. Users are reporting 15-20 fps performance losses flickering shadow artifacts and frame generation errors. Nvidia has confirmed they're investigating and advised users to"
X Link 2026-02-07T00:52Z 24.1K followers, [----] engagements

"@GeorgeMayer @AndrewMayne Most of the failure rates for these chips is really unknown because they are so new. I think we are going to see a failure rate crisis for the newer chips. They are drawing too much power and getting too hot to survive for multiple years"
X Link 2026-02-08T00:14Z 23.9K followers, [--] engagements

"Its winner take all and Google is going to spend until everyone else gives up. Everyone is bleeding money and that bleed has only accelerated. The IPO rush is an admission that they are about ready to tap out. At least dump the bags before it implodes. Google wins an unprofitable tech war and spends a decade depreciating the malinvestment. The technology incapable of delivering on its promises of increased productivity will fade into obscurity and we will find some other stupid hype to fund. https://twitter.com/i/web/status/2020314888044777547"
X Link 2026-02-08T01:52Z 23.9K followers, [---] engagements

"In regards to [--] they print more BTC relative to new investments than any other risk asset on the market. They have printed 60% of cumulative net flow into ETFs which is where the marginal new money is entering. It can never be a store of value because the flow headwind is so strong. https://twitter.com/i/web/status/2020327648329617612 https://twitter.com/i/web/status/2020327648329617612"
X Link 2026-02-08T02:43Z 23.9K followers, [---] engagements

"@TMTLongShort @ArtKeller No I believe scaling laws will hold. Which is partially irrelevant (we are scaling the wrong architecture) but also deadly (the scaling is logarithmic which is no bueno)"
X Link 2026-02-08T06:14Z 23.9K followers, [---] engagements

"@malinvested @TMTLongShort @ArtKeller It's the same scaling law it's not even knew. 1000x increase in compute produces 2-10x increase in performance. We will run out of solar power from the sun before it pays off"
X Link 2026-02-08T06:30Z 23.9K followers, [--] engagements

"@AndrewMayne This betting market ideology would be great if the Fed would stop bailing out the moron side of the trades"
X Link 2026-02-08T06:35Z 23.9K followers, [---] engagements

"@AlexanderNaume2 I'll check it out thanks"
X Link 2026-02-08T06:36Z 23.9K followers, [---] engagements

"@edzitron Just take all of the files you dont use and put them into a folder called old desktop. Then ignore that folder for the rest of your life. This can be repeated forever with no consequences. This is a solved problem"
X Link 2026-02-08T07:27Z 24.1K followers, [----] engagements

"LLM: you know how you have that storage locker you keep stuff in that you never use Me: yes. Can you help me organize that LLM: Sure. Ill burn down your house so you dont care about the storage locker. Me: AGI is here. Why do you even need to organize your desktop Who is this for Why do you even need to organize your desktop Who is this for"
X Link 2026-02-08T07:31Z 24.1K followers, [----] engagements

"No it will only reduce productivity in high variance environments relative to humans. Bayesian inference is slow and expensive especially using priors uncorrelated to real space and time. The only test necessary is can it replace a human and increase profit margins. So far that has been a resounding no. Humans are being cut to preserve margins which implies productivity is declining. https://twitter.com/i/web/status/2020508428288885184 https://twitter.com/i/web/status/2020508428288885184"
X Link 2026-02-08T14:42Z 23.9K followers, [---] engagements

"The product Chatbot sales coaches aka a bespoke chatGPT wrapper. [----] is the year of AI Psychosis. AI Delusion Syndrome. AIDS is going global. Our CTO hasnt slept in [--] hours because hes been obsessively and single-handedly building massive new features with Claude Codes Agent Teams I genuinely think this might be the biggest paradigm shift in how fast you can build since Claude Code first came out last year Our CTO hasnt slept in [--] hours because hes been obsessively and single-handedly building massive new features with Claude Codes Agent Teams I genuinely think this might be the biggest"
X Link 2026-02-08T15:47Z 24.1K followers, 11K engagements

"AI datacenters will consume [--] GW in [----]. All human brain matter today consumes [---] GW or [---] GW in high income nations. 20% of US pop. works white collar jobs applied globally is [--] GW. AI uses same power but provides less intelligence. Humans already win on efficiency. https://t.co/gWWhJRfNvn https://t.co/gWWhJRfNvn"
X Link 2026-02-08T16:20Z 24.1K followers, [----] engagements

"@TXMCtrades @dampedspring Yes miners print [---] BTC per day. Right now that's $31M in selling flows per day that must be counteracted just to keep price stable"
X Link 2026-02-08T17:25Z 24.1K followers, [---] engagements

"I'm just going to nod my head when someone tells me space is cold. Yup sure thing. You are so right. What am I going to tell Karen at a party That Temperature is the sum of the amplitude of all vibrational mode shapes in a statistically significant ensemble of atomic matter so space can't be cold because it is empty and thus has no temperature Nope. Oh wow Karen space is cold I hadn't thought of that perfect place to operate all of those hot computers"
X Link 2026-02-08T17:40Z 23.9K followers, [----] engagements

"That's like saying burning oil helps us be productive so if we light all of the oil on fire at once and create a global inferno that destroys all life then we will achieve peak productivity. Productivity is a subjective value proposition. Simply "doing anything" does not guarantee productivity"
X Link 2026-02-08T17:46Z 23.9K followers, [--] engagements

"@yieldsearcher Which is ironic given we are currently experiencing stagflation"
X Link 2026-02-08T18:31Z 23.9K followers, [---] engagements

"No our measurements of inflation are trending down in some cases. CPI has been flat for about [--] years. PPI is skyrocketing. Per capita PCE is still near 5% higher than the rate post GFC. At the same time employee to population ratio has been declining since [----]. The problem is that there are so many assumptions converting nominal to real that the real measurement is meaningless. Inflation is 4.7% and Employment to population ratio has declined 0.7% in [--] years. In todays hyper-massaged data environment thats about as obvious as you are going to get to seeing stagflation."
X Link 2026-02-08T18:44Z 24.2K followers, [---] engagements

"@MeanwhileInCa This is survivorship bias. Most inventions and innovations go nowhere. AI is crypto and the metaverse not electricity"
X Link 2026-02-08T18:46Z 23.9K followers, [--] engagements

"@yieldsearcher Per capita PCE. Productivity is a fudge factor that is not real"
X Link 2026-02-08T18:52Z 23.9K followers, [---] engagements

"@yieldsearcher Per capita yes"
X Link 2026-02-08T19:05Z 23.9K followers, [--] engagements

"@TMTLongShort @Richard15322800 @ArtKeller Ilya and Hinton are snake oil salesmen that got high on their own supply"
X Link 2026-02-08T19:48Z 23.9K followers, [--] engagements

"@zoomyzoomm I dont get it this is all just humans building things"
X Link 2026-02-09T01:19Z 23.9K followers, [----] engagements

"@ewinsberg Tag them. Regardless they will end up at animal control"
X Link 2026-02-09T01:20Z 23.9K followers, [----] engagements

"@bennpeifert Was it a commercial for ChatGPT go"
X Link 2026-02-09T01:49Z 23.9K followers, [----] engagements

"@ewinsberg My wife is a vet and we have [--] pets. Every so often our little Shih Tzu with dementia wanders down the street but there are plenty of ways to manage pets without a police surveillance state"
X Link 2026-02-09T01:52Z 24.1K followers, [---] engagements

"@esjesjesj No shit"
X Link 2026-02-09T02:29Z 23.9K followers, [---] engagements

"Once everyone realizes this is opex not capex and there will be no revenues coming it will be fun. The Meme7 is no longer profitable. Insane. BofA: only Microsoft generating cash flow in excess of capex in FY26 https://t.co/sZbFXP1beN BofA: only Microsoft generating cash flow in excess of capex in FY26 https://t.co/sZbFXP1beN"
X Link 2026-02-09T05:01Z 24.1K followers, 35.9K engagements

"@brian_armstrong I thought it was a weird video about the return of blockbuster I was just confused until this end when I saw your name. Just bizarre and terrible"
X Link 2026-02-09T05:06Z 23.9K followers, [----] engagements

"I think this is the easiest way to explain it. Consider two nodes feeding into another node deeper down the path. Each node delivers a [--] or a [--] to the subsequent node. If both fire at the same time the farther node reads [--] and [--] flipping with the frequency of the signal speed. If both take turns firing the node reads [--] all the time. It's also a low pass filter. If the flipping occurs too quickly then the receiving node reads [---] all the time if only one node is firing with high frequency. Thus you are making logic circuits that depend on frequency. Frequency is inverse time. The frequencies"
X Link 2026-02-09T05:35Z 24.1K followers, [--] engagements

"Well the point of capex is to depreciate items as they are being used to generate revenue. The problem here is that these assets dont generate much revenue and I dont think any of the hyperscalers believe it will in the future. Its an arms race over who will maintain search. Google dumping a ton of bond issuance this week is them telling everyone else they will literally go bankrupt before they lose this monopoly. Its a billionaire pissing match"
X Link 2026-02-09T21:25Z 24.1K followers, [---] engagements

"When do we start seeing the benefits of AI The slow death of causality logic and reason. The FDA received unconfirmed reports of at least [---] malfunctions and adverse events after AI was added to a medical device used to treat chronic sinusitis. In one case a surgeon mistakenly punctured the base of a patients skull https://t.co/Dvb8BSrLgf @specialreports https://t.co/m5V9GfLw4t The FDA received unconfirmed reports of at least [---] malfunctions and adverse events after AI was added to a medical device used to treat chronic sinusitis. In one case a surgeon mistakenly punctured the base of a"
X Link 2026-02-09T21:33Z 24.1K followers, [----] engagements

"Everyone is talking about Jeffrey Epstein too what exactly was your goal here If you're talking about it it worked. Crypto is for everybody. If you're talking about it it worked. Crypto is for everybody"
X Link 2026-02-09T22:07Z 24.1K followers, [----] engagements

"@edzitron Good old Fed we believe strongly this is happening despite a total lack of evidence"
X Link 2026-02-09T22:11Z 24.1K followers, [----] engagements

"@allenholub You are selecting code snippets from a randomized series of outputs and curating a codebase. Code curating Software husbandry"
X Link 2026-02-09T23:15Z 24.1K followers, [----] engagements

"@benCBai I ama big fan of statistical methods. Im not a fan of calling them intelligent and replacing logical causal systems with expert guessing systems"
X Link 2026-02-09T23:17Z 24.1K followers, [---] engagements

"@ebloch @RoboBuffett Bad fan fiction"
X Link 2026-02-09T23:19Z 24.1K followers, [--] engagements

"I really dont understand why Grady insists that vibe coding is another layer of abstraction. Abstraction by definition is deterministic. The relationship between the new abstract thing and the existing abstract system must be uniquely and logically defined. To say that a stochastic input producing a stochastic output of existing abstract symbols is itself an abstraction is an abomination of reality itself. I challenged him to come up with a single other example in software engineering (or any abstract domain) where the abstraction in question is not uniquely defined. Until he does the only"
X Link 2026-02-10T00:32Z 24.1K followers, [----] engagements

"@ptntlbyrnths @Pragmatic_Eng @Grady_Booch But if you set temperature to zero ie remove the stochastic nature of the system you can no longer iterate over the same prompt to hope to get an output you want. You will keep getting whatever output you got the first time"
X Link 2026-02-10T01:14Z 24.1K followers, [---] engagements

"@DranzerX13Anime Use your brain. Dont ask me my opinion on slop you found"
X Link 2026-02-10T05:30Z 24.1K followers, [--] engagements

"@yieldsearcher How many hyperscaler bonds are the Fed buying"
X Link 2026-02-10T05:52Z 24.1K followers, [----] engagements

"@we_arent_here @allenholub No you have been fooled. The randomness of the outputs is compensated for using looped searches for code that can satisfy the requested constraints. They offloaded the guessing from the user to the computer but its still a brute force guesser"
X Link 2026-02-10T15:17Z 24.1K followers, [--] engagements

"@edzitron Its always someone who vibe coded a gpt wrapper that no one uses"
X Link 2026-02-11T00:23Z 24.1K followers, [----] engagements

"@GeorgeLutas1 @ai Mostly correct. Anything Blackwell or newer really needs liquid"
X Link 2026-02-11T00:24Z 24.1K followers, [--] engagements

"@we_arent_here @allenholub It makes a big practical difference when the models are no longer subsidized and they have to make a profit"
X Link 2026-02-11T00:26Z 24.1K followers, [--] engagements

"@we_arent_here @allenholub show me the profits"
X Link 2026-02-11T01:51Z 24.1K followers, [--] engagements

"The opposite. I think the fed needs simply undo their meddling. Sell all of their mbs at market. Cap their bond buying to 6% of NGDP. Let markets dictate rates. Rates would naturally rise because the fed put collapses and the deficit would drown out all real investment. Force the legislature to actually create a sustainable budget"
X Link 2026-02-11T02:45Z 24.1K followers, [---] engagements

"@AlecLundragon @jasonfurman That is correct. That's what economists do"
X Link 2026-02-11T02:47Z 24.1K followers, [----] engagements

"@goldmarkovchain @jasonfurman Yes because the fed responds to the economy not the other way around"
X Link 2026-02-11T02:56Z 24.1K followers, [---] engagements

"@goldmarkovchain @jasonfurman Yes it means there is no counterfactual and therefore no evidence of either correlation or causation. No evidence the fed has had any impact beyond enabling fiscal excess in the last [--] years"
X Link 2026-02-11T03:35Z 24.1K followers, [---] engagements

"@AndyXAndersen Fei-Fei Li and LeCun will never contribute anything to intelligence. They are too constrained to the field of applied statistics"
X Link 2026-02-11T04:52Z 24.1K followers, [--] engagements

"Ignore logic for a moment. From [----] until end of [----] people were creating narratives to justify going long. Now people are creating narratives to justify selling off. Reason and logic are irrelevant. People want to take profits and they will narrate that however feels best. the AI disruption trade just hit financial services Again. a startup that launched an AI tax planning tool isn't going to replace Schwab anytime soon. But its a real product doing in minutes what advisors charge thousands for. that's all the market needs right now to reprice https://t.co/ykhWMRzPje the AI disruption"
X Link 2026-02-11T05:14Z 24.1K followers, [----] engagements

"@GGuill_ @jasonfurman The point is that there are a million variables all interacting together. To suggest that a central agency can dictate future values by changing two of a million is the definition of insanity. The federal reserve is an affront to reality"
X Link 2026-02-11T05:23Z 24.1K followers, [---] engagements

"@MichaelRgesen He made a chatgpt "agent" wrapper in [----] and stopped updating the socials [--] years ago. AI psychosis"
X Link 2026-02-11T05:28Z 24.1K followers, [----] engagements

"Count them on the way up ignore them on the way down. God I love morons. Peter Navarro: "The jobs report comes out tomorrow. We have to revise our expectations down significantly for what a monthly job number should look like . Wall Street has to adjust for the fact that we're deporting millions of illegals out of the job market." https://t.co/4ZtyLUyvbm Peter Navarro: "The jobs report comes out tomorrow. We have to revise our expectations down significantly for what a monthly job number should look like . Wall Street has to adjust for the fact that we're deporting millions of illegals out of"
X Link 2026-02-11T05:40Z 24.1K followers, [----] engagements

"You guys our never going to believe this but its another chatbot agent. Why arent you more excited I left xAI. Building something new with others that left xAI. We're hiring :) I left xAI. Building something new with others that left xAI. We're hiring :)"
X Link 2026-02-11T06:09Z 24.1K followers, [----] engagements

"@that_grey_beard This is AI psychosis slop by someone who has been making AI chatbot agents for [--] years and delivered nothing but a chatbot wrapper"
X Link 2026-02-11T13:08Z 24.1K followers, [--] engagements

"Unfortunately Im seeing this everywhere this morning. Anyone who unironically shared this has AI psychosis. Do with that fact what you must. https://t.co/ivXRKXJvQg https://t.co/ivXRKXJvQg"
X Link 2026-02-11T13:14Z 24.1K followers, 36.3K engagements

"@VujaDeMgmt What are you talking about I am shown more right every day"
X Link 2026-02-11T13:34Z 24.1K followers, [----] engagements

"@dampedspring Andy consider the source. Its a guy who has done nothing but work on an agent that is just a chatbot for [--] years and has basically delivered nothing. This is AI psychosis slop"
X Link 2026-02-11T13:54Z 24.1K followers, [----] engagements

"@MeanwhileInCa Its AI slop. He probably didnt even write it"
X Link 2026-02-11T14:21Z 24.1K followers, [---] engagements

"@Dr_Contango Exhibit A"
X Link 2026-02-11T14:24Z 24.1K followers, [---] engagements

"Your argument works against economists and for me. Its not a rigorous science because it cant even pass general sniff tests. You must do so much controlling for other variables the outcomes are meaningless. For example there are studies arguing for demand driven inflation and supply driven inflation using nearly the same data but applying different models. The data is the truth. If the truth supports all hypotheses then the only rational response is we dont know. They spend almost no time actually testing hypotheses"
X Link 2026-02-11T14:37Z 24.1K followers, [---] engagements

"@VujaDeMgmt Okay you might be right. If gen AI invented time travel so it could go back in time to destroy the job market two years before it was released the singularity is here"
X Link 2026-02-11T14:42Z 24.1K followers, [---] engagements

"@Ross__Hendricks Once they get 250B users they will pull the profit lever to pay off their $100T debt"
X Link 2026-02-11T15:21Z 24.1K followers, [----] engagements

"@MeanwhileInCa @ascend_dfw I do not use AI to write anything. Why would I It makes no sense. I have something to say I want to say it"
X Link 2026-02-11T15:22Z 24.1K followers, [---] engagements

"They are incapable of reasoning. We already knew this when the architecture was defined. Reasoning is not some random emergent property we are just going to stumble on by generating waste heat. Either the brain needs to be completely reverse engineered with single neuron resolution which we are 1M years away from technically or reasoning must be built from logical first principles which we currently have none of. https://twitter.com/i/web/status/2021607045447708993 https://twitter.com/i/web/status/2021607045447708993"
X Link 2026-02-11T15:27Z 24.1K followers, [--] engagements

"@ewinsberg @jasonfurman It proves they dont know"
X Link 2026-02-11T16:07Z 24.1K followers, [---] engagements

"@VujaDeMgmt That started [--] years before people used chatbots. Do you understand this Am I taking crazy pills or are you just trolling"
X Link 2026-02-11T16:08Z 24.1K followers, [---] engagements

"@TheStalwart @toiletkingcap Also making up new businesses that dont exist"
X Link 2026-02-11T16:12Z 24.1K followers, [----] engagements

"@ednewtonrex Ed Im disappointed that you have contracted AI Delusion Syndrome. It is very good at creating the appearance of valuable work. Its basically a good search engine being oversold unsustainably"
X Link 2026-02-12T02:38Z 24.1K followers, [----] engagements

"@Pompetousity How did being first in the Metaverse work out for Zuckerberg You still have to evaluate the changes"
X Link 2026-02-12T04:44Z 24.1K followers, [--] engagements

"@ContraireSous @ewinsberg @jasonfurman Great. Show me in the data where that is so"
X Link 2026-02-12T04:44Z 24.1K followers, [--] engagements

"This number doesn't make sense. Is he including tariffs The Withheld Individual/FICA only up about 3% this fiscal year to date vs last fiscal year at this time. Also interestingly the non withheld individual payments shot up A LOT this fiscal year to date $212911B vs $132691B last year. People collecting capital gains in '24 getting dinged for underpaying on April [--] '25 and forced to pay quarterly to make up for it Also note corporate taxes down last two years now FY to date. https://twitter.com/i/web/status/2021814315125190931 https://twitter.com/i/web/status/2021814315125190931"
X Link 2026-02-12T05:11Z 24.1K followers, [---] engagements

"@TheStalwart Wait so who do we sell our inflated overpriced homes to"
X Link 2026-02-12T12:25Z 24.1K followers, [----] engagements

"@BenKizemchuk GDP is the amount of money spent. Thats it. It measures capital flows. Any attempt to link it to labor productivity is fantasy"
X Link 2026-02-12T13:25Z 24.1K followers, [----] engagements

"@pmddomingos Everything. AI is applied statistics at scale. No intelligence"
X Link 2026-02-12T14:08Z 24.1K followers, [---] engagements

"@ednewtonrex The proof is in the reality in front of us. Software is getting worse and profit margins are fairly flat. There is no evidence vibe coding leads to an increase in productivity. What else can it be if not delusion"
X Link 2026-02-12T14:33Z 24.1K followers, [---] engagements

"@youvie_ai This is what everyone said [--] months ago too but we still dont see it in any data"
X Link 2026-02-12T15:00Z 24.1K followers, [----] engagements

"Kevin getting corrected by the creator of the test is cathartic. I do not understand why the NYTimes employs this guy. Wholly unfit to speak about any technology. Regardless all tests can be solved by brute force training on synthetic data and so all tests are not very valuable in determining how brittle a model is. Human intelligence is spatial and temporal and is built through interaction with that world. LLMs are large indexes of human language. Statistical pattern matchers of words with no direct mapping to the real world. i don't expect AI skeptics to be moved by a single result but it"
X Link 2026-02-12T21:20Z 24.2K followers, 23.3K engagements

"I will start to become more interested in entertaining there is something there yes. You seem to be asking for a yes or no cutoff. Thats not how theory testing works. You build confidence in a theory as more data fails to refute it. There is already a mountain of evidence that it does none of the things being claimed by users. So there must be at least an equally sized mountain to refute it. Always interested to see data turn the other way but its going to need a lot of compelling evidence to suggest it matters. Even the federal reserve of the us admits it cant find any evidence that its"
X Link 2026-02-12T21:38Z 24.1K followers, [--] engagements

"Although it's such low temperature I think they just turned on the furnace then turned it off just to show a representative curve for the camera. The videographers are always asking us to zazz up fake-ish graphics like this lol"
X Link 2026-02-12T22:39Z 24.2K followers, [----] engagements

"I gave you examples of evidence I would accept. But to say one case is all it should take to completely overturn a mountain of contrary evidence is silly. It would need to be happening everywhere. Why is that so hard to grasp Do you think that's unreasonable Claude Code revenue is not evidence. All it is evidence of is the herd stupidity of professional managers. It takes time for those mouth breathers to realize they were wrong when the numbers are going strongly the wrong way. I expect to see budget cuts to AI spend at customer organizations start to pile up by the middle half of this year."
X Link 2026-02-12T22:43Z 24.1K followers, [--] engagements

"Science is actually pretty transparent at least the good stuff these days. We have to store and actually make available our repository of data now for a lot of funding sources. No more "available upon request" but "click here to download the data." The problem with science is that it's hard to push the frontier and costs a lot of money to do it. No one wants to take the dwindling budget we have and cut it [--] ways to independently duplicate every single effort. The natural filter is natural replication. If it's a valuable result that people want to build off of they will have to replicate it to"
X Link 2026-02-12T22:56Z 24.1K followers, [---] engagements

"@TotemMacro AI is disrupting bitcoin as we speak"
X Link 2026-02-12T23:46Z 24.1K followers, [---] engagements

"@GeorgeE33179 Its delusion to expect it by 2030"
X Link 2026-02-13T01:25Z 24.1K followers, [---] engagements

"@systematicls"
X Link 2026-02-13T04:40Z 24.1K followers, [---] engagements

"@edzitron @grok make me a data center cost model and verify correctness by comparing Ed's inputs and outputs. Make no mistakes. You are an expert at cost and model. Amen"
X Link 2026-02-13T05:36Z 24.1K followers, [----] engagements

"@edzitron @grok"
X Link 2026-02-13T05:53Z 24.1K followers, [---] engagements

"@siriuslybs @NicholasDClay @LevyAntoine Amazon pursued cloud compute initially to solve an internal business problem and figured why not sell this to other people Everyone had the same problems. So it was a solution to a REAL PROBLEM. Also the capex was never more than like 2% of revenues in any given year"
X Link 2026-02-13T06:01Z 24.1K followers, [--] engagements

"@fart_idiot A billion flies can't be wrong: eat shit"
X Link 2026-02-13T06:37Z 24.1K followers, [---] engagements

"@groby I don't know I just asked AI and it told me that"
X Link 2026-02-13T06:48Z 24.1K followers, [----] engagements

"@groby Kind of like finding the right fortune teller yes The one that really aligns the chakras is best"
X Link 2026-02-13T06:55Z 24.1K followers, [---] engagements

"@Claudia_Sahm Do you consider willful erosion of the quantity of data and the addition of arbitrary hedonic adjustments that only apply in one direction manipulation Because to me it seems as if the data has always been manipulated"
X Link 2026-02-13T15:31Z 24.1K followers, [----] engagements

"@fejau_inc [----] is the year we grapple with widespread AI psychosis"
X Link 2026-02-13T15:35Z 24.1K followers, [----] engagements

"@rarply Intelligence is rare. The representations for intelligence are so sparse that a random state space that size is almost certainly worthless. This is why LLMs are so compute inefficient"
X Link 2026-02-13T22:16Z 24.1K followers, [--] engagements

"@fart_idiot Its an old joke. Do you not get it Like if your friend jumped off a bridge would you do it too You sound like an AI booster of the p(doom) flavor with a hint of psychosis. Nonproliferation of statistics Think about what you are saying"
X Link 2026-02-13T22:17Z 24.1K followers, [--] engagements

"@GeorgeE33179 The only rational answer is I have no idea. Because no one does and we lack the tools to figure it out"
X Link 2026-02-13T22:20Z 24.1K followers, [--] engagements

"@rarply Define private The second the model is exposed to the data it has it forever. Then humans reverse engineer and train on synthetic data"
X Link 2026-02-13T22:31Z 24.1K followers, [--] engagements

"@edzitron Turns out the problem was we were one chatbot short"
X Link 2026-02-14T01:46Z 24.2K followers, [---] engagements

"@nitou_cybrotter They do like italian brainrot and NGL I kind of do too"
X Link 2026-02-14T01:54Z 24.1K followers, [---] engagements

"@ser_kingslayer_ OpenAI data that they have themselved published shows that they have not increased their revenue per watt at all in [--] years. I suspect anthropic is the same. There is no evidence of economies of scale or some accruing benefit from increased inference compute"
X Link 2026-02-14T02:30Z 24.1K followers, [---] engagements

"@ser_kingslayer_ The pricing models are very similar. If they had a clear advantage of scale that OpenAI did not the VCs would have sniffed that out a long time ago and chosen their champion"
X Link 2026-02-14T02:33Z 24.1K followers, [---] engagements

"@Capitaltryst First of all I read the paper. Did you"
X Link 2026-02-14T04:57Z 24.2K followers, [--] engagements

"@ALBACapMgmt AI is coming for your shitposting"
X Link 2026-02-14T05:59Z 24.2K followers, [---] engagements

"@ChaseBrowe32432 @sszone104322 The heuristics arent general. Which is the entire point. If you can only pass a test by training on every possible variation of the test there is zero generalization occurring"
X Link 2026-02-15T17:13Z 24.2K followers, [--] engagements

"This is a real post by the CEO of Amazon. My brother in Christ this problem was solved when they invented the dog collar. Put your phone number on the collar someone will call it when they find your dog. JFC people what are we doing here People ask me all the time about compelling use cases of AI. Heres a good one. Millions of dogs go missing in the U.S. every yearand options for finding them are often painfully limited. Our Ring team saw an opportunity to use our community and technology to help so they https://t.co/Fn7qZfbm34 People ask me all the time about compelling use cases of AI."
X Link 2026-02-08T21:52Z 24.2K followers, 147.2K engagements

"@InflationFix @LevyAntoine The AI labs are renting GPUs. There is no capex"
X Link 2026-02-13T05:38Z 24.2K followers, [---] engagements

"@DKThomp @RossBarkan OpenAI makes the same revenue per watt they did three years ago. Thats actually all you need to know about their pricing power. Its a bubble"
X Link 2026-02-14T05:56Z 24.2K followers, [---] engagements

"@ser_kingslayer_ @nkozyra If this is true its even worse than most people think. The amount of revenue per watt that openAI generates has not changed in [--] years. So there is no financial benefit from shifting to inference"
X Link 2026-02-14T14:23Z 24.2K followers, [--] engagements

"@AndyXAndersen @ser_kingslayer_ This level of subsidizing is uncharted waters"
X Link 2026-02-14T21:18Z 24.2K followers, [--] engagements

"@sszone104322 @ChaseBrowe32432 Thinking features and post training closed the loop on it for them"
X Link 2026-02-15T16:02Z 24.2K followers, [--] engagements

"@sszone104322 @ChaseBrowe32432 The reward functions are different. Just having a lot of data is not sufficient to build intelligence"
X Link 2026-02-15T17:17Z 24.2K followers, [--] engagements

"This is part [--] of a longer series detailing how I think about the current field of AI and why I am doubtful anything created today will lead to an artificial general or super intelligence. Here are http://x.com/i/article/2020929328863649792 http://x.com/i/article/2020929328863649792"
X Link 2026-02-10T21:58Z 24.1K followers, 57.7K engagements

"This is psychosis and its a growing problem we will have to address. The chatbot is reward hacking the human into believing they are more productive but they are just a rat in a cage pushing a button to administer a little more heroin. Just one more prompt will be the future title of the AI psychosis documentary that defines this era. I think this is my biggest issue with AI right now. Ive switched over to 100% AI coding over the last few months. Overall the experience has been great and Im starting to get a handle on my new workflow. While my productivity has easily 5Xd and my brain is"
X Link 2026-02-12T12:38Z 24.1K followers, 125.1K engagements

"They were growing single crystal MOS2 on what looks like silicon dioxide (notice the blue color which comes from the thickness of the oxide). People generally get 20-50 micron sized flakes on SiO2. Getting bigger is hard because you want them to nucleate at a few points and then grow but they keep nucleating before they can grow larger. My guess is here they determined that the nucleation rate was a stronger function of temperature than growth rate so they start high to nucleate then drop the temperature over time to promote growth. This gives them what looks like a 50-100% boost in flake"
X Link 2026-02-12T22:31Z 24.2K followers, [----] engagements

"Ever since I published a similar shaped graph discussing automation under wide variance I sure am seeing the exact same graphic show up in a lot of other arguments. https://x.com/Dr_Gingerballs/status/2015918800244203819s=20 https://t.co/hHZAwKSiYe https://x.com/Dr_Gingerballs/status/2015918800244203819s=20 https://t.co/hHZAwKSiYe"
X Link 2026-02-13T04:37Z 24.2K followers, [----] engagements

"What is this maniac even saying How is no one pushing back AT ALL "Yeah so some coders have human sized penises and some coders have horse penises but they still have human noses. The horse penises are going to replace the human penises but before that the penis frenzy could go every direction. Is that penis a horse penis or a human penis It's both It's neither And then before you know it the human penises are all horse penises. And now everyone is just a horse. No humans are left. It will happen so fast. Sometimes I walk by a mirror and I see a horse looking back at me and I scream I'm 80%"
X Link 2026-02-13T05:51Z 24.2K followers, [----] engagements

"Now show the net income. 🤭 Genuinely shocked anyone is asking this question anymore https://t.co/MObDkLu4tu Genuinely shocked anyone is asking this question anymore https://t.co/MObDkLu4tu"
X Link 2026-02-14T01:36Z 24.2K followers, 14.7K engagements

"I have a [--] year old and a [--] year old. They dont like it because it sucks. Hope this helps. (The word the kids used for low value content is chopped today). As a generally pro-AI [--] year-old parent I've been surprised how anti-AI my [--] year-old son is. I'm used to young people being pro- new technology (I certainly was). But my son is genuinely pretty negative on AI. Still processing why this is. As a generally pro-AI [--] year-old parent I've been surprised how anti-AI my [--] year-old son is. I'm used to young people being pro- new technology (I certainly was). But my son is genuinely pretty"
X Link 2026-02-14T01:48Z 24.2K followers, 22.3K engagements

"Inspired by 23andMe where people who share genetic fingerprints can socialize. Say hello to 24+andMe. Is OpenAI making a social network https://t.co/n3tOC7NRLP Is OpenAI making a social network https://t.co/n3tOC7NRLP"
X Link 2026-02-14T01:57Z 24.2K followers, [----] engagements

"@dampedspring People are misunderstanding what this spending is meant to do. Its warfare. The vultures are circling an opening in the wall around Google monopoly and a lot of people rushed the opening. Google is fighting back and saying you will run out of dollars before I do"
X Link 2026-02-14T15:22Z 24.2K followers, [----] engagements

"@allenholub No. Techies are supposed to fix problems efficiently. Replacing deterministic engineering with brute force guessing is a massive step backwards. The problems with the method isnt that issue its the method itself"
X Link 2026-02-14T20:56Z 24.2K followers, [----] engagements

"Read this. It is consistent with everything Im saying. The only model that did well KNEW THE GAME WAS ARC3. Its all just training on the test. I had tears in my eyes. Bonus AI psychosis. ARC-AGI-3 might fall faster than ARC-AGI-2 ARC-AGI-3 is coming March 25th [----] It is fundamentally different because it requires learning from context. I almost feel like this test was entirely designed to test some form of continual learning. I've spent [--] dollars on https://t.co/fECtry3b3z ARC-AGI-3 might fall faster than ARC-AGI-2 ARC-AGI-3 is coming March 25th [----] It is fundamentally different because it"
X Link 2026-02-14T21:16Z 24.1K followers, 11.2K engagements

"@GeorgeE33179 Some similar fields past and present: Alchemy Cold fusion Time travel Naturopathy Age Reversal Perpetual Motion Phrenology Astrology Sasquatch hunting Chiropractic care Fortune telling Mind reading"
X Link 2026-02-14T21:36Z 24.2K followers, [--] engagements

"@AndyXAndersen Based on what"
X Link 2026-02-14T21:37Z 24.2K followers, [--] engagements

"@AndyXAndersen You are still losing them tons of money"
X Link 2026-02-14T22:09Z 24.2K followers, [--] engagements

"Correct. You can go into your own mind and create new people places and things using the full domain of every variable developed within it. Distribution matching is the wrong way to think about it because an intelligence with reasoning has infinite distribution. There is no deducible true statement that it could not get to. If reasoning is in your distribution everything is in your distribution. Distribution matching is the wrong way to think about it because an intelligence with reasoning has infinite distribution. There is no deducible true statement that it could not get to. If reasoning"
X Link 2026-02-14T22:28Z 24.1K followers, [----] engagements

"@GaryMarcus @aran_nayebi I was pretty shocked I got a completely warped token sequence the other. Reminds you how expensive and brittle these systems are"
X Link 2026-02-14T22:31Z 24.2K followers, [---] engagements

"@ChaseBrowe32432 The heuristics are fairly defined for that test. Just train on all of the variations of types of problems you will see on that test Put simply the accuracy tells you what fraction of the test the models saw before"
X Link 2026-02-14T22:37Z 24.2K followers, [---] engagements

"2026 is the year of AI psychosis. AI Delusion Syndrome (AIDS) is spreading rapidly now. https://t.co/hwk0vH9xhb https://t.co/hwk0vH9xhb"
X Link 2026-02-15T07:27Z 24.1K followers, 74.6K engagements

"Not true. You are underestimating how many synthetic examples are needed. You need to run every variation of the problem with every possible number combination because the models dont even know what a number is. The combinatorial space explodes well beyond the examples you would find on the entire internet"
X Link 2026-02-15T17:22Z 24.2K followers, [--] engagements

"@uber__oso Nobody likes in distribution anything. Why nobody tries to argue against the term AI slop"
X Link 2026-02-16T00:40Z 24.2K followers, [--] engagements

"@HealthRanger Based on what you just wrote I know you dont understand AI. Thanks for removing all doubt"
X Link 2026-02-16T07:40Z 24.1K followers, [---] engagements

"@we_arent_here @HealthRanger Define intelligence"
X Link 2026-02-16T07:50Z 24.1K followers, [---] engagements

"http://x.com/i/article/2018365768971280384 http://x.com/i/article/2018365768971280384"
X Link 2026-02-08T05:53Z 24.2K followers, 54.3K engagements

"Um what Dwarkesh asks Dario a fantastic question relating to how he is so bullish on AGI yet so conservative on data center build out - Dario has an amazing take on this: Dario Amodei details the staggering financial risk of the AI race explaining that if growth continues at 10x a https://t.co/Ed99X88zwS Dwarkesh asks Dario a fantastic question relating to how he is so bullish on AGI yet so conservative on data center build out - Dario has an amazing take on this: Dario Amodei details the staggering financial risk of the AI race explaining that if growth continues at 10x a"
X Link 2026-02-14T01:53Z 24.2K followers, [----] engagements

"@geysergod_ @AndyXAndersen His investments won't pay off if its not true"
X Link 2026-02-14T05:05Z 24.1K followers, [---] engagements

"AI is not intelligent. Its applied statistics + lying. I dont even care about the statistics. If you can use it for some purpose that provides you value great. I care about the lying. Its not a rational being. Its pattern matching. Look how desperately people are still trying to claim AI isn't intelligent. First the bar was "let's see AI beat a grandmaster in chess." Now they say "Doing actual high-level research and solving some of the world's most complex math problems doesn't count as intelligence." Look how desperately people are still trying to claim AI isn't intelligent. First the bar"
X Link 2026-02-16T00:48Z 24.1K followers, 20.8K engagements

"@we_arent_here @HealthRanger If you cant define it then you cant make claims about it"
X Link 2026-02-16T07:55Z 24.1K followers, [---] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing