@slow_developer Haider.
Haider. posts on X about ai, open ai, agi, anthropic the most. They currently have [------] followers and [---] posts still getting attention that total [-------] engagements in the last [--] hours.
Engagements: [-------] #
- [--] Week [---------] -6.80%
- [--] Month [---------] -54%
- [--] Months [----------] +92%
- [--] Year [-----------] +82%
Mentions: [---] #
- [--] Week [---] +31%
- [--] Month [---] +22%
- [--] Months [-----] +93%
- [--] Year [-----] +128%
Followers: [------] #
- [--] Week [------] +0.53%
- [--] Month [------] +2.80%
- [--] Months [------] +29%
- [--] Year [------] +78%
CreatorRank: [-------] #
Social Influence
Social category influence technology brands 39.91% stocks #3555 celebrities 5.5% finance 2.29% social networks 1.83% countries 1.38% gaming 0.46% products 0.46%
Social topic influence ai #2342, open ai #281, agi #22, anthropic #670, model #1406, in the 5.96%, systems #741, ceo #460, $googl 5.5%, xai #78
Top accounts mentioned or mentioned by @grok @newaiworld_ @barrakali @william_r2rclub @tmallard @joewilliams010 @kyletranai @kazuakat @softsyncdxb @justmichael_80 @bygregorr @andygrossberg @martinszerment @airevolthq @langrisser4o @navneet_rabdiya @wgreklek @alphamfpefm @p33rl3ss @thejennyai
Top assets mentioned Alphabet Inc Class A (GOOGL) Microsoft Corp. (MSFT)
Top Social Posts
Top posts by engagements in the last [--] hours
"xAI should've called Grok [---] "Grok 4.5" because the jump is huge Grok [---] uses large-scale RL and new agentic reward models to boost: - creative writing - personal-tone control - emotional intelligence and the biggest win which few notice is the big drop in hallucinations: Grok 4: 12.09% Grok 4.1: 4.22% that's roughly 3x fewer hallucinations about a two-thirds reduction"
X Link 2025-11-18T07:10Z 62.5K followers, 15.5M engagements
"Anthropic Felix Rieseberg: "we're still very early in AI models and products will improve fast likely faster than most expect even if you disagree picture 10x better models and tools to understand why frontier labs are fixated on 2026" you can see the trend locking in now. even if AI progress stopped now better tools around today's abilities would still push us slowly toward AGI https://twitter.com/i/web/status/2020864009725952120 https://twitter.com/i/web/status/2020864009725952120"
X Link 2026-02-09T14:15Z 62.5K followers, [----] engagements
"gpt-5.2 is already better than me in most areas except creative and outside-the-box thinking it can be a bit narrow. i'm still better at finding new ways to look at a problem but once the problem is clearly defined the model usually does a better job applying the solution"
X Link 2026-02-15T01:45Z 62.5K followers, [----] engagements
"Anthropic CEO Dario Amodei AI coding capabilities will reach a "very serious" level by the end of [----] and may match the best human coders by late [----] I feel this threatening because we are the ones building it"
X Link 2025-03-01T19:40Z 62.5K followers, 246.1K engagements
"Terence Tao says that for the next 10-20 years humans and AI will possess complementary strengths AI can synthesize a million papers and test every idea inside them. Humans can look at [--] examples and say "I see the pattern now" "they can try to fake it but they're very inefficient at it" https://twitter.com/i/web/status/2007633081034006904 https://twitter.com/i/web/status/2007633081034006904"
X Link 2026-01-04T02:00Z 62.5K followers, 618.9K engagements
"2026 is going to be a real turning point for AI both in the technology itself and in how society reacts to it. the recent progress has been very fast and by the end of [----] it will be hard for anyone to still pretend nothing big is happening"
X Link 2026-02-15T08:50Z 62.5K followers, [----] engagements
"Sam Altman says even as AI becomes unbelievably intelligent it will not become the center of the human story "we're wired to care about people not machines" And we will continue to value human achievements and connections far more than machine performance"
X Link 2025-09-27T17:00Z 62.4K followers, 229.9K engagements
"Anthropic co-founder Jack Clark: by summer [----] the AI economy may move so fast that people using frontier systems feel like they live in a parallel world to everyone else most of the real activity will happen invisibly in digital AI-to-AI spaces with only surface signs showing up in everyday life (datacenters compute/power constraints and the startup ecosystem) https://twitter.com/i/web/status/2003903502012256332 https://twitter.com/i/web/status/2003903502012256332"
X Link 2025-12-24T19:00Z 62.4K followers, 280.1K engagements
"Physicist Brian Cox says consciousness is the most complex emergent phenomenon we know but it still comes from underlying laws In principle a powerful enough computer could model the brain AI and AGI raise this question: "we still don't know how consciousness emerges from the brain" https://twitter.com/i/web/status/2005277556027904315 https://twitter.com/i/web/status/2005277556027904315"
X Link 2025-12-28T14:00Z 62.2K followers, 104.5K engagements
"Perplexity CEO Aravind Srinivas says the biggest threat to data centers is intelligence that runs locally on your device If models run on your own chip they can watch your workflow and adapt via "test time training" without data ever leaving your computer "you own it. it's your brain""
X Link 2026-01-01T14:00Z 62.4K followers, 846.2K engagements
"Jensen Huang says the industry is shifting from generative AI to agentic AI The next major step is fusing public cloud frontier models with customized open-source systems running on enterprise servers But the end goal is physical: "moving intelligence from the cloud into industrial AI factories and robotics" https://twitter.com/i/web/status/2007451886094360688 https://twitter.com/i/web/status/2007451886094360688"
X Link 2026-01-03T14:00Z 62.4K followers, 50.3K engagements
"the next possible model release is grok [----] i'm expecting it to be right up there with the current top frontier models not gemini for a while perhaps gemini [--] GA claude and gpt-5.2 checkpoint might be next but probably not claude [--] or gpt-6 more likely it'll be smaller updates https://twitter.com/i/web/status/2009373294441398464 https://twitter.com/i/web/status/2009373294441398464"
X Link 2026-01-08T21:15Z 62.2K followers, 10.9K engagements
"Terence Tao says the current business model of big AI companies relies on hype to justify their massive valuations The situation is similar to the dot-com boom where the tech was real but the early promises were inflated "hopefully we'll reach a more realistic equilibrium in a year or two" https://twitter.com/i/web/status/2009445018763673960 https://twitter.com/i/web/status/2009445018763673960"
X Link 2026-01-09T02:00Z 62.1K followers, 87.6K engagements
"the reason we mostly see openai models solving math proofs is because gpt reasoning models have very strong RL which helps a lot with math. gpt-5.2 pro is even better than gpt-5.2 thinking for this kind of work and the extra thinking budget usually adds a lot more value"
X Link 2026-01-12T05:20Z 62.4K followers, [----] engagements
"dario seems very confident that recursive self-improvement could happen by [----] he hinted claude [---] could help build claude [--] i don't think this is unrealistic opus [---] is already impressive so moving to [--] or [---] could mean parts of claude [--] are built using earlier versions"
X Link 2026-01-21T16:15Z 62.2K followers, 10K engagements
"both openai and anthropic researchers have confirmed that: codex and claude code can write all of their code. i've said this recently and am getting a lot of hate from delusional people who still want to write code from scratch the era of writing code is over and reading code has begun https://twitter.com/i/web/status/2015421899581612436 https://twitter.com/i/web/status/2015421899581612436"
X Link 2026-01-25T13:50Z 62.4K followers, 66.2K engagements
"yann lecun left meta and chose WAR i think he's right. we keep making the same mistake: we see "superhuman" results in a narrow area and immediately start yelling AGI but it also shows how fast we normalize things like show today's models to someone 5-10 years ago and they'd call it superintelligence https://twitter.com/i/web/status/2015603093526745454 https://twitter.com/i/web/status/2015603093526745454"
X Link 2026-01-26T01:50Z 62.1K followers, 75.8K engagements
"Demis Hassabis says AI will disrupt entry-level jobs first so mastering these tools now beats a traditional internship The capability overhang of these tools offers massive creative leverage But this is only the next five years "once agi arrives we'd be in uncharted territory" https://twitter.com/i/web/status/2017024963828449651 https://twitter.com/i/web/status/2017024963828449651"
X Link 2026-01-30T00:00Z 62.4K followers, 15.6K engagements
"Anthropic Logan Graham: "our view is that in [----] anthropic crossing a threshold where self-improving systems become possible" i'm not sure about the claim but i do think labs already have a basic "AI researcher agent" that can brute-force algorithm and training improvements https://twitter.com/i/web/status/2017583652352495908 https://twitter.com/i/web/status/2017583652352495908"
X Link 2026-01-31T13:00Z 62.1K followers, [----] engagements
"i'm really looking forward to gpt-5.3 sam asked for feedback on gpt-5.3 last week which makes it feel like a release will not be that far but if he's still asking what ppl want improved it probably means the model is still in training or at least not fully finished yet"
X Link 2026-01-31T16:30Z 62.2K followers, 12.5K engagements
"it's weird how creative writing is often treated as unimportant gpt-5.1 and gemini [--] are pretty decent but nowhere near as good as gpt-4.5 opus [---] is still the best right now which is funny because it's the one model that never even marketed itself around creative writing"
X Link 2026-02-01T01:30Z 62.1K followers, [----] engagements
"ar the rate AI is moving GTA [--] might feel outdated just a couple of years after it comes out world models like Genie and spatial intelligence work like World Labs will push games far beyond what we have today so GTA [--] should end up being the last "traditional" GTA game https://twitter.com/i/web/status/2017910106051293609 https://twitter.com/i/web/status/2017910106051293609"
X Link 2026-02-01T10:37Z 62.2K followers, 84K engagements
"there's a lot of disagreement about what "AGI" means one common idea is simple: an AI that can do most digital economic work basically anything a remote worker does i think openai definition is close to that an AI capable of performing many economically valuable tasks"
X Link 2026-02-01T15:40Z 62.2K followers, [----] engagements
"xAI is on the path to AGI google is on the path to AGI openai is on the path to AGI anthropic is on the path to AGI open-source labs are on the path to AGI unfortunately company loyalty has gotten very strong lately instead of supporting the progress being made"
X Link 2026-02-02T04:00Z 62.4K followers, [----] engagements
"i hope and expect gpt-5.3 to be an absolute beast in reasoning and speed gpt-5.2 is already very strong though its higher-reasoning variants are a bit slow it was released only three weeks after gpt-5.1 still i think we'll get GPT-5.2-codex-max first"
X Link 2026-02-02T06:30Z 62.2K followers, 15.1K engagements
"Ex-OpenAI Researcher Jerry Tworek: Coding agents are the next level of abstraction in programming a new language with different semantics "soon almost no one will type code directly" The focus will shift to ensuring software reliability when humans are neither writing nor reading it"
X Link 2026-02-02T09:00Z 62.2K followers, 13.5K engagements
"many people think AGI can solve everything but that's not really true take a real-world problem that smart people could solve in theory and ask: "why hasn't it been solved" most of the time the answer isn't intelligence it's cost coordination time risk or constraints https://twitter.com/i/web/status/2018285771422318649 https://twitter.com/i/web/status/2018285771422318649"
X Link 2026-02-02T11:30Z 62.2K followers, [----] engagements
"Geoffrey Hinton says people who call AI just a stochastic parrot are wrong The models don't store text; they convert words into complex sets of features They predict the next word by processing these features in context not by mindlessly recombining language from the web"
X Link 2026-02-02T18:35Z 62.2K followers, 90.1K engagements
"codex app is a big launch codex [---] is already ahead of claude code if you're not a vibe coder so if sonnet [--] (per rumors) can't beat it that's a big win for openai and if codex [---] ends up faster and better than [---] they've basically locked it in i love this competition"
X Link 2026-02-02T21:00Z 62.2K followers, 32.4K engagements
"gemini [--] pro is generally smarter than opus [---] in overall intelligence opus [---] is better when it comes to programming deepmind focuses on building a strong general reasoning and search model claude models are mainly optimized for programming and coding tasks"
X Link 2026-02-02T23:25Z 62.4K followers, 34.3K engagements
"seedance [---] pro is another noticeable jump for image-to-video it can turn a static image into a cinematic-looking clip with natural motion and believable expressions also supports native audio with accurate lip-sync including multilingual dialogue and multiple speakers overall the output looks polished and stable for narration ads and other pro use cases https://twitter.com/i/web/status/2018502196732612705 https://twitter.com/i/web/status/2018502196732612705"
X Link 2026-02-03T01:50Z 62.4K followers, [----] engagements
"i have a feeling gpt-5.3 is around the corner today while testing gpt-5.2 warmer personality i got a pop-up saying i'm helping build a new version of GPT and asked which response i prefer side-by-side i've seen these pop-ups before but they usually don't mention a new version https://twitter.com/i/web/status/2018573919347954113 https://twitter.com/i/web/status/2018573919347954113"
X Link 2026-02-03T06:35Z 62.4K followers, 18.5K engagements
"i don't know but svg tests don't really excite me opus [---] is already very good at front-end design so as gemini [--] what i really want is the next claude version to handle tough backend work like hard database migrations and messy real-world integrations where it has to reason through edge cases https://twitter.com/i/web/status/2018590560966369387 https://twitter.com/i/web/status/2018590560966369387"
X Link 2026-02-03T07:41Z 62.2K followers, [----] engagements
"Yann LeCun says basing agentic systems on the current LLM paradigm is a recipe for disaster Intelligent behavior requires a world model to predict the consequences of actions which LLMs do not have "the basic architecture is not there" https://twitter.com/i/web/status/2018610422237323643 https://twitter.com/i/web/status/2018610422237323643"
X Link 2026-02-03T09:00Z 62.5K followers, 61.8K engagements
"qwen-coder-qoder is now live this custom model is built for Qoder's agent workflow and i'm genuinely impressed by how it understands the entire repo and keeps pushing until the code actually works very refreshing to see this level of autonomy in a coding tool excited to see how it handles complex tasks Qwen-Coder-Qoder is live Weve launched a customized model built on Alibabas Qwen-Coder fine-tuned via large-scale RL specifically for the Qoder. Learn more: https://t.co/xwL3oF6tpq https://t.co/8SjcVw5peH Qwen-Coder-Qoder is live Weve launched a customized model built on Alibabas Qwen-Coder"
X Link 2026-02-03T16:15Z 62.4K followers, [----] engagements
"Mo Gawdat says we're heading toward an economy where machines make everything and none of us have jobs Since 62% of the US GDP relies on consumption the economy will collapse unless society pays everyone to live "basically everyone's going to UBI other than very few" https://twitter.com/i/web/status/2018755112970441051 https://twitter.com/i/web/status/2018755112970441051"
X Link 2026-02-03T18:35Z 62.2K followers, 24.9K engagements
"sam hinted again that gpt-5.3 isn't far away there are multiple rumors that gpt-5.3 will be much more intelligent possibly the "garlic" model with "IMO gold-level" reasoning 400k context window and up to 128k output. if that's true my guess is those limits will be for thinking or pro users only https://twitter.com/i/web/status/2018864584359059737 https://twitter.com/i/web/status/2018864584359059737"
X Link 2026-02-04T01:50Z 62.4K followers, 17.1K engagements
"Geoffrey Hinton says we can guess a few years ahead but [--] years out Nobody knows The best education now is to learn to think independently and understand STEM so you can understand what AI is doing AI lags in physical dexterity "so being a plumber will outlast being a lawyer" https://twitter.com/i/web/status/2018972806373941284 https://twitter.com/i/web/status/2018972806373941284"
X Link 2026-02-04T09:00Z 62.2K followers, 27.6K engagements
"i'm not sure where the sonnet [--] launched on feb [--] hype came from even if "claude-sonnet-5@2026020" is real the date in the name is probably not the release day opus [---] had "20251101" in its name but launched on nov [--] but let's see because the hype is back again today https://twitter.com/i/web/status/2019010546134303156 https://twitter.com/i/web/status/2019010546134303156"
X Link 2026-02-04T11:30Z 62.2K followers, 12.2K engagements
"sam altman: "things are about to move quite fast and we will be working with extremely powerful models soon" - openai is hiring a head of preparedness - google is hiring a post-AGI economist no wonder their internal models are months ahead of what will be released publicly https://twitter.com/i/web/status/2019045778321457167 https://twitter.com/i/web/status/2019045778321457167"
X Link 2026-02-04T13:50Z 62.4K followers, 17.4K engagements
"sam openly takes shots at anthropic and i think he's right. claude on the free plan is one of the most limited chatbots out there and that's with only around [--] million monthly users chatgpt has about [---] million weekly active users with around 95% of them on the free tier 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"
X Link 2026-02-04T21:00Z 62.4K followers, 49.9K engagements
"gpt-5.2 is clearly the smartest model METR latest evals show: gpt-5.2-high at a 50% time-horizon of [---] hours 80% time-horizon of [--] minutes which means it can complete a software task a human estimates at [---] hours on 50% of tries and a 55-minute task on 80% of tries and keep in mind this is just "high" reasoning effort not "xhigh" https://twitter.com/i/web/status/2019408166128349244 https://twitter.com/i/web/status/2019408166128349244"
X Link 2026-02-05T13:50Z 62.4K followers, [----] engagements
"have you noticed that nearly all AI haters seem to have incredibly low IQ i've been debating AI doubters lately and once you counter their nonsensical arguments with evidence they get angry and start personal attacks job anxiety plays a big part. when something feels threatening denial is the easiest defense basic education gaps matter too many adults struggle with simple math so they can't grasp AI and assume it's fake or overhyped. because of this AI is creating a divide people who learn and adapt will move forward while others keep denying it. so arguing with those who refuse to learn"
X Link 2026-02-05T16:15Z 62.2K followers, 12.5K engagements
"what a day for the so-called AI bubble openai launched gpt-5.3 codex probably the best agentic model so far anthropic launched opus [---] with only small improvements but already close to saturating arc-agi-2 and suddenly google is out of the conversation that's how cruel this game is https://twitter.com/i/web/status/2019520154011373756 https://twitter.com/i/web/status/2019520154011373756"
X Link 2026-02-05T21:15Z 62.4K followers, 60.8K engagements
"in my view both dario and demis have shared their paths to AGI: - dario is betting on superhuman coding and math models that can improve themselves and accelerate progress - demis agrees but also emphasizes continuous learning multimodal AI and a deeper understanding of the real world i guess it will be a hybrid of the two https://twitter.com/i/web/status/2019594393062080709 https://twitter.com/i/web/status/2019594393062080709"
X Link 2026-02-06T02:10Z 62.4K followers, [----] engagements
"if we dropped gpt-5.3 codex or opus [---] into [----] we'd probably have called it AGI or at least clearly beyond human level in a few domains but the definition keeps shifting and the goalposts keep moving so yeah the "AGI by 2025" crowd wasn't totally wrong https://twitter.com/i/web/status/2019667373859696774 https://twitter.com/i/web/status/2019667373859696774"
X Link 2026-02-06T07:00Z 62.4K followers, [----] engagements
"gpt-5.3 codex was used to create itself claude builds opus [---] i'm not a fan of the "singularity" sci-fi label but this is basically the same hypothesis predicted years ago models are getting better at helping build better models. i think the main bottleneck now is compute this still takes months because you need tens of thousands of GPUs running nonstop https://twitter.com/i/web/status/2019737838489939982 https://twitter.com/i/web/status/2019737838489939982"
X Link 2026-02-06T11:40Z 62.2K followers, 13.3K engagements
"Jensen Huang says the next frontier is physical AI systems that understand the physical world and causality A child naturally recognizes the basic physics of tipping over dominoes (gravity mass contact) LLMs have no idea "we have to create a new type of physical AI" https://twitter.com/i/web/status/2019810820457111903 https://twitter.com/i/web/status/2019810820457111903"
X Link 2026-02-06T16:30Z 62.4K followers, 33.1K engagements
"hollywood is kind of cooked even if you believe the "AI is a bubble" crowd (i don't) realistic video models in 5-10 years will likely be able to make full movies from just a few prompts i think the real limit won't be the tech it'll be IP laws if those get reformed AI video could fully take off Kling [---] is truly "one giant leap for AI video generation" Check out this amazing mockumentary from Kling AI Creative Partner Simon Meyer https://t.co/Iyw919s6OJ Kling [---] is truly "one giant leap for AI video generation" Check out this amazing mockumentary from Kling AI Creative Partner Simon Meyer"
X Link 2026-02-06T21:25Z 62.4K followers, [----] engagements
"AI is testing AI now from the opus [---] system card: "anthropic increasingly relied on claude models to help run and debug its own evaluation setup under time pressure" i think this shows how fast these systems are getting where even testing them now needs their help https://twitter.com/i/web/status/2019955522980733128 https://twitter.com/i/web/status/2019955522980733128"
X Link 2026-02-07T02:05Z 62.4K followers, [----] engagements
"gemini [--] is still a great general model i use codex for coding and i use gemini in the chat interface because i often prefer it over chatgpt/claude since google doesn't financially depend on keeping hype alive they can realistically go a while without releasing a new model"
X Link 2026-02-07T04:30Z 62.1K followers, 10.8K engagements
"- openai is hiring a head of preparedness - anthropic is already a safety-first lab i think both labs' internal worries are about self-improving AI dario has hinted at this lately. they don't want to lose control but competition will force them to give models more freedom anyway https://twitter.com/i/web/status/2020029765399580676 https://twitter.com/i/web/status/2020029765399580676"
X Link 2026-02-07T07:00Z 62.4K followers, 16.7K engagements
"in jan [----] only a small part of the code was written by AI by winter [----] most code was written by AI by the end of this year almost all code will be written by AI if that isn't acceleration i don't know what is it took from [----] to [----] to move from 0% to 100% of code being written by AI https://twitter.com/i/web/status/2020102742765920754 https://twitter.com/i/web/status/2020102742765920754"
X Link 2026-02-07T11:50Z 62.4K followers, [----] engagements
"opus [---] looks like a real jump in math the model has basically caught up with openai best models and is even ahead of gemini [--] on most difficult math benchmarks on the hardest level (Tier 4) opus [---] solved [--] out of [--] in comparison opus [---] solved only [--] out of 48"
X Link 2026-02-07T14:10Z 62.4K followers, 10.4K engagements
"gpt-5.3 codex seems as smart as gpt-5.2 but faster and more focused on coding so output quality stays high it still sometimes stops too early and assumes the problem is simpler than it is but it doesn't get stuck and if you keep going it usually finds the fix the speed makes the trade-off worth it"
X Link 2026-02-07T16:40Z 62.1K followers, [----] engagements
"Sam Altman says the magic of the last [--] years is simple: more resources = better models on a log scale People keep predicting it will plateau or slow down. They've been wrong every time "compute power is the new oil" sometimes you spend on data sometimes on compute sometimes on new ideas https://twitter.com/i/web/status/2020210957058801674 https://twitter.com/i/web/status/2020210957058801674"
X Link 2026-02-07T19:00Z 62.2K followers, 35.4K engagements
"gpt-4o was the GOAT for conversations but after gpt-5 launched it's become terrible and often praises way too much when a system approves everything you do it doesn't help you improve. it makes you comfortable with average results and disconnects you from reality bring on gpt-5.3"
X Link 2026-02-07T21:30Z 62.2K followers, [----] engagements
"2026 will be the year agent hierarchies go mainstream if they work as expected they could enable a hard take-off: agents building better agents. we're already seeing early signs and within [--] months we may get self-contained coding + science systems that can handle almost anything"
X Link 2026-02-07T23:50Z 62.1K followers, [----] engagements
"genie [--] launched last week but it feels like nothing because it was already demoed months ago and it's locked behind the $250/m plan with no pay-as-you-go api that basically shuts out most normal users i get why they did it (server load would be crazy) but it's still annoying"
X Link 2026-02-08T02:15Z 62.4K followers, [----] engagements
"opus [---] is very much comparable to gpt-5.2 high which is impressive i'm also interested to see how gpt-5.3 codex performs when its API is released in my experience [---] codex felt lazier than the default [---] and the lower effort was clearly visible in the build quality"
X Link 2026-02-08T04:40Z 62.2K followers, [----] engagements
"@ForrestPKnight the comparison itself is wrong opus [---] is an overall general-purpose frontier model codex [---] is a coding-focused model/agent a better comparison would be the updated claude code vs codex 5.3"
X Link 2026-02-08T05:27Z 62.1K followers, [----] engagements
"gemini [--] flash is so good because it's basically a distilled gemini [--] pro and google pushed it with extra RL for agentic tool use. if they did that for flash they've likely been doing the same for the full pro model too and with a jan [----] cutoff and slower releases they've probably been cooking bigger upgrades for a while https://twitter.com/i/web/status/2020392152619184530 https://twitter.com/i/web/status/2020392152619184530"
X Link 2026-02-08T07:00Z 62.4K followers, 27.4K engagements
"Elon Musk says that by the end of this year AI will solve digital human emulation A "digital remote worker" that can do anything involving moving electrons or improving human productivity That's the ceiling before robots "once you have physical robots the limit is physics itself" https://twitter.com/i/web/status/2020427381874532644 https://twitter.com/i/web/status/2020427381874532644"
X Link 2026-02-08T09:20Z 62.3K followers, 19.4K engagements
"i'm really hoping we get gpt-5.3 next week codex [---] has already been released and sam has been teasing it nonstop it seems likely they might launch around the same time when gpt-4o/gpt-5 is deprecated as they shift users to the new models gpt-5.3 is going to be a banger"
X Link 2026-02-08T14:10Z 62.4K followers, 26.1K engagements
"claude code is great in many ways but it often moves too fast and even skips files codex on the other hand reads every file it can find before taking action this helps with tasks that need full context. CC works best for quick tasks and brainstorming while codex is better for implementation"
X Link 2026-02-08T16:40Z 62.2K followers, [----] engagements
"Geoffrey Hinton says LLMs are immortal As long as their weights are saved the machine can be destroyed and the same system can be brought back later on new hardware Many churches claim resurrection but this is something we can actually do "only for digital things""
X Link 2026-02-08T19:00Z 62.4K followers, 25.6K engagements
"i cringe when people say "can't wait for the AI bubble to pop" most of them haven't really used it or the ones who did probably tried it back in [----] when chatgpt was still early they don't realize how much it has improved. i think it's not the AI bubble popping it's their own little bubble"
X Link 2026-02-08T21:30Z 62.4K followers, [----] engagements
"opus [---] is a big improvement in general intelligence at least a half-step change but definitely not incremental and i don't think these models will dominate discussions for months since even better releases are expected soon maybe for vibe coding it could be true"
X Link 2026-02-08T23:50Z 62.2K followers, [----] engagements
"we were supposed to hit a wall AI progress was hitting a wall after gpt-5 release but here we are global AI compute capacity is roughly doubling every [--] months METR's earlier estimate was [--] months and now it seems closer to 4-5 so if AI "hits a wall" it's bad news for the wall"
X Link 2026-02-09T02:15Z 62.2K followers, 10.4K engagements
"xAI hasn't been in the spotlight lately but i'm excited for grok [--] it's training on the colossus [--] supercomputer reportedly the biggest and most powerful AI supercluster right now using more power than major cities this should be a xAI biggest release coming in q2 2026"
X Link 2026-02-09T04:40Z 62.4K followers, 16.9K engagements
"the idea that codex is now helping to create new versions of codex claude is building the next claude model is exciting but also a bit scary. it makes you wonder how close we are to gpt-5.4 or sonnet [--] so yes a feedback loop exists but it's not (yet) the autonomous "self-improving itself" stage https://twitter.com/i/web/status/2020754537615900761 https://twitter.com/i/web/status/2020754537615900761"
X Link 2026-02-09T07:00Z 62.4K followers, [----] engagements
"Sam Altman says OpenAI is in a hurricane of public anxiety and media scrutiny The outside world fixates on rumors often 99% divorced from reality. Internally it's calmer and focused on compute and the next model "it is less chaotic internally than i think you would imagine" https://twitter.com/i/web/status/2020789770352578654 https://twitter.com/i/web/status/2020789770352578654"
X Link 2026-02-09T09:20Z 62.4K followers, 41.5K engagements
"it makes sense for openai to release gpt-5.3 this week codex [---] already dropped last week and is getting a lot of attention because it's already incredible and since they're deprecating gpt-4o and gpt-5 on feb [--] it looks like they're pushing everyone toward the newer models"
X Link 2026-02-09T16:35Z 62.4K followers, [----] engagements
"Elon Musk says corporations that are purely AI and robotics will vastly outperform any with humans in the loop Computers replaced entire skyscrapers of human "calculators" now a laptop with a spreadsheet does more than [--] floors of people ever could "this shift will happen very quickly" https://twitter.com/i/web/status/2020935733356151201 https://twitter.com/i/web/status/2020935733356151201"
X Link 2026-02-09T19:00Z 62.4K followers, 18K engagements
"ray kurzweil predicted AGI in [----] and the singularity around [----] i mostly agree with him right now machines are still limited but if progress continues that'll change dramatically in the next [--] years still humans also struggle to grasp exponential growth: we expect too much too soon and too little later https://twitter.com/i/web/status/2020973480234381515 https://twitter.com/i/web/status/2020973480234381515"
X Link 2026-02-09T21:30Z 62.4K followers, 10.4K engagements
"so Meta is building a new LLM codenamed "Avocado" even before fine-tuning or alignment the model's raw pretraining results are already beating today's top open-source models even if that's true Meta will likely need at least [--] more months to catch up on features compared to the other labs https://twitter.com/i/web/status/2021043945061658770 https://twitter.com/i/web/status/2021043945061658770"
X Link 2026-02-10T02:10Z 62.2K followers, [----] engagements
"@TFWNicholson i don't think so right now it looks more like a human + AI feedback loop that is accelerating progress not AI running away on its own"
X Link 2026-02-10T06:21Z 62.2K followers, [---] engagements
"@paradite_ the difference is that openai actively takes feedback to improve the product even from the CEO at anthropic it seems like only boris and tariq take feedback and not very regularly also codex is open-source so we can publicly report bugs and request features"
X Link 2026-02-10T06:44Z 62.2K followers, [---] engagements
"i'm super hyped for the gpt-5.3 release the information reported: - gpt-5.3 is expected to be a fully new pre-trained model - gpt-5.2 was described as an early checkpoint of the full "garlic" model not the final version there's a lot of conflicting info some say it's a garlic checkpoint but a lot of people think gpt-5.3 is the full garlic release"
X Link 2026-02-10T06:55Z 62.2K followers, 20.7K engagements
"@Alibaba_Qwen qwen just can't stop winning"
X Link 2026-02-10T08:31Z 62.2K followers, [----] engagements
"it takes extreme ignorance to believe LLMs are plateauing or even close to it people have been saying "LLMs have plateaued" since [----] but i've learned to ignore them and look at where the money is going and what engineers in the field are actually building not the skeptics"
X Link 2026-02-10T11:50Z 62.2K followers, [----] engagements
"gpt-5.3 codex xhigh achieves 90% on the Next.js eval up from 40% with gpt-5.2 codex the tests measure how well coding agents handle Next.js tasks such as code generation and migrations primarily through success rate i'm not surprised since codex has been the real deal for backend coding https://twitter.com/i/web/status/2021225138742596019 https://twitter.com/i/web/status/2021225138742596019"
X Link 2026-02-10T14:10Z 62.4K followers, 17K engagements
"what's going on at xAI one of their co-founders has left the company i would imagine people are feeling a bit burned out from the intense work and power issues during training and with all the recent xAI business moves i wouldn't be surprised if more people leave and new hires come in I resigned from xAI today. This company - and the family we became - will stay with me forever. I will deeply miss the people the warrooms and all those battles we have fought together. It's time for my next chapter. It is an era with full possibilities: a small team armed I resigned from xAI today. This company"
X Link 2026-02-10T16:40Z 62.4K followers, 15.6K engagements
"Fei-Fei Li says building world models requires moving beyond internet data to massive real-world capture and simulated data similar to how self-driving car companies work this combines real-world data collection with synthetic data generation There is a flywheel: the models will generate simulation data for robotics https://twitter.com/i/web/status/2021298119493800394 https://twitter.com/i/web/status/2021298119493800394"
X Link 2026-02-10T19:00Z 62.5K followers, 11.7K engagements
"i hate all this nonsense talk about openai introducing ads why are people suddenly acting like ads are bad now especially once AI starts doing them just stop who really cares they're just ads. advertising has been a basic part of business since the printing press"
X Link 2026-02-10T21:30Z 62.4K followers, [----] engagements
"gemini fast growth makes sense because google has integrated it across its products claude growth is more interesting because it is driven mostly by developers grok growth is the wild card chatgpt has been growing even faster since december open source is nearing closed-source levels https://twitter.com/i/web/status/2021444081394569683 https://twitter.com/i/web/status/2021444081394569683"
X Link 2026-02-11T04:40Z 62.2K followers, [----] engagements
"i'm really curious about a "gpt-5o" version in the gpt-5 series there are rumors that a more advanced voice mode is coming. since gpt-4o will be deprecated soon the current advanced voice that runs on it may also be replaced or upgraded but the next model update is gpt-5.3"
X Link 2026-02-11T07:00Z 62.4K followers, [----] engagements
"Mathematician Terence Tao says we're still jury-rigging AI into our workflows It's not yet like working with a human collaborator we've spent centuries perfecting how we work with other humans "but we haven't figured out the best way to work with AI yet" https://twitter.com/i/web/status/2021514545571803243 https://twitter.com/i/web/status/2021514545571803243"
X Link 2026-02-11T09:20Z 62.4K followers, 11.7K engagements
"opus [---] is clearly better than [---] at getting UI layouts right on the first attempt especially with complex layouts that opus [---] would need 2-3 iterations on we've been running it through claude code and it understands component spacing instead of just throwing flexbox at everything https://twitter.com/i/web/status/2021549778899960150 https://twitter.com/i/web/status/2021549778899960150"
X Link 2026-02-11T11:40Z 62.4K followers, [----] engagements
"@marksuman why are low-IQ people so obsessed with payouts instead of responding to the argument how much do you think X pays me for this tweet $20 $50"
X Link 2026-02-11T15:24Z 62.3K followers, [----] engagements
"@abram2k1 @marksuman ok i apologise for misunderstanding @marksuman"
X Link 2026-02-11T16:15Z 62.4K followers, [---] engagements
"i think we're going to get another paradigm shift that will curve the exponential even steeper. agency and RSI can greatly boost AI capability and most AI labs are very focused on them right now so i'd be surprised if there isn't a breakthrough this year even if it doesn't fully solve it https://twitter.com/i/web/status/2021622761349615636 https://twitter.com/i/web/status/2021622761349615636"
X Link 2026-02-11T16:30Z 62.4K followers, [----] engagements
"googles been moving noticeably slower on model releases than anthropic and openai gemini [--] pro preview dropped in nov but there still isn't a stable release which makes it feel stuck in "preview limbo" anthropic meanwhile has been shipping fast opus [---] / [---] came out within roughly two months https://twitter.com/i/web/status/2021695739538420022 https://twitter.com/i/web/status/2021695739538420022"
X Link 2026-02-11T21:20Z 62.4K followers, 14.7K engagements
"the AI race isn't slowing down and a bubble burst doesn't look likely it's simple: if the US slows down china can take the lead and if AI companies run into money trouble they'll point to fast-moving chinese open-source models as proof that the government needs to step in Introducing GLM-5: From Vibe Coding to Agentic Engineering GLM-5 is built for complex systems engineering and long-horizon agentic tasks. Compared to GLM-4.5 it scales from 355B params (32B active) to 744B (40B active) with pre-training data growing from 23T to 28.5T tokens. https://t.co/uGYQUjIbbs Introducing GLM-5: From"
X Link 2026-02-11T23:50Z 62.4K followers, [----] engagements
"there's no reason to believe intelligence has a fixed limit and with an exponential curve even a small early lead becomes a big advantage. for example the METR benchmark shows that the time it takes models to handle tasks is twice as long as it is getting shorter"
X Link 2026-02-12T04:40Z 62.4K followers, [----] engagements
"claude code might be 3-4x faster but i barely use it because codex [---] high does the job better most of the time and why sit around waiting for an agent to build a feature even if you wait its still 10100x faster than us [----] was the year writing code fundamentally changed"
X Link 2026-02-12T07:00Z 62.4K followers, [----] engagements
"@ai_for_success my longest codex (gpt-5.2 codex) run was [--] hours [--] minutes it handled a full end-to-end task: planning refactoring the codebase updating database schemas and migrations running the full test suite and improving the code etc until everything worked properly"
X Link 2026-02-12T08:09Z 62.4K followers, [---] engagements
"we always focus too much on the next big breakthrough but we've barely scratched the surface of multimodal integration feedback loops memory and agency even if LLMs plateau the most interesting progress will come from the scaffolding around them the orchestrator layer"
X Link 2026-02-12T11:50Z 62.4K followers, [----] engagements
"@sama codex mob app"
X Link 2026-02-12T16:33Z 62.4K followers, [----] engagements
"codex is for serious developers who focus on real work instead of hype claude code is popular with vibe coders who like to show off quick one-shot results and cool demos both are good depending on your goals"
X Link 2026-02-12T16:40Z 62.4K followers, 11.4K engagements
"@andygrossberg yup codex extension. i'm mostly using that compared to the cli"
X Link 2026-02-12T16:57Z 62.4K followers, [---] engagements
"@OpenAINewsroom good. replace it with gpt-5.3 now"
X Link 2026-02-12T17:02Z 62.4K followers, [----] engagements
"@_nik1st my experience with claude models especially on large codebases and backend work is not great it kinda try to please you and say your code is good to go even when it's clearly wrong"
X Link 2026-02-12T17:08Z 62.4K followers, [---] engagements
"@Angaisb_ did they release the gpt-5.2-codex-max"
X Link 2026-02-12T17:18Z 62.4K followers, [----] engagements
"@Funnybuilds i think it depends what we mean by iterative is it iterative because it asks good follow-ups and refines well or because it needs more back-and-forth to land the result"
X Link 2026-02-12T17:26Z 62.4K followers, [---] engagements
"the codex team literally can't sleep or eat openai just dropped a new model: "gpt-5.3-codex-spark" it's basically an ultra-fast 'small' codex model powered by cerebras hardware. but it's only available for pro users and api users for now GPT-5.3-Codex-Spark is now in research preview. You can just build thingsfaster. https://t.co/85LzDOgcQj GPT-5.3-Codex-Spark is now in research preview. You can just build thingsfaster. https://t.co/85LzDOgcQj"
X Link 2026-02-12T18:26Z 62.4K followers, 39.2K engagements
"@paulhshort for a small set of users. not sure who they are from the blog: in addition we are making Codex-Spark available in the API for a small set of design partners"
X Link 2026-02-12T18:35Z 62.4K followers, [----] engagements
"Terence Tao says humans are bad at specifying goals and AI is good at fulfilling them If you ask it for proof it might cheat by inventing axioms or changing definitions to satisfy the request So we must define what we want "because often it's not just the solution it's understanding" https://twitter.com/i/web/status/2022050577564549224 https://twitter.com/i/web/status/2022050577564549224"
X Link 2026-02-12T20:50Z 62.4K followers, 10.1K engagements
"i think it's clear we entered a slow takeoff last summer even if we haven't closed the loop yet AI models are speeding up AI research at an accelerating rate and if progress stopped today it would still take many years to fully understand and use what we've already built"
X Link 2026-02-13T06:30Z 62.5K followers, [----] engagements
"Microsoft CTO Kevin Scott says scaling laws were clear [--] years ago AI would be powerful general and platform-like The only question was the speed but it's happening faster than expected "we are not yet at the point of diminishing marginal returns" on the infrastructure powering this change https://twitter.com/i/web/status/2022267003868787009 https://twitter.com/i/web/status/2022267003868787009"
X Link 2026-02-13T11:10Z 62.4K followers, 13.8K engagements
"it looks like gpt-5.3 post-training is taking longer than expected openai is probably spending extra time refining guardrails to make everything work properly not a fan of guardrails but they help prevent misuse and keep companies and legal teams comfortable so i don't think it's slow progress https://twitter.com/i/web/status/2022304752697000217 https://twitter.com/i/web/status/2022304752697000217"
X Link 2026-02-13T13:40Z 62.4K followers, 14.1K engagements
"google cheif scientist jeff dean recently explained the gemini next model cycle: train a massive frontier model first then distill its capabilities into smaller faster versions that's how the new-gen 'flash' can end up beating the previous-gen 'pro' it's like you inherit the frontier brain but run it cheaper + faster https://twitter.com/i/web/status/2022376475190579695 https://twitter.com/i/web/status/2022376475190579695"
X Link 2026-02-13T18:25Z 62.4K followers, 19K engagements
"we are in a very unusual situation AI can handle days weeks or even months of work but it still needs humans involved many people misunderstand this. once models are persistent keep learning and have real tool access everything changes they might already be beyond human-level in some ways"
X Link 2026-02-14T01:45Z 62.4K followers, [----] engagements
"learning to code has always been valuable. getting a CS degree hasn't always been necessary. a CS degree isn't mainly about learning to code. it's more about the foundations and bigger ideas behind computing don't confuse the two"
X Link 2026-02-14T06:30Z 62.4K followers, [----] engagements
"@chrisparkX what about the schedule post's chris"
X Link 2026-02-14T09:46Z 62.4K followers, [---] engagements
"Google should release multiple model versions for Gemini [--] like OpenAI does: - gpt-5 chat - gpt-5 mini - gpt-5 pro - gpt-5 high - gpt-5 codex each version would serve a specific purpose. limiting everything to just one or two models isn't ideal unless they let users pick the reasoning level themselves"
X Link 2025-11-15T14:40Z 62.5K followers, 22.8K engagements
"since the latest METR result with opus [---] we've entered an era of almost-vertical progress all it will take is a few more jumps like this and we could be heading into software-on-demand and RSI progress hasn't just matched METR's earlier 7-month doubling time; it's beaten it now it looks like 4-5 months. incredible"
X Link 2025-12-21T10:00Z 62.5K followers, 28.2K engagements
"this year will be huge no doubt new data centers are coming online and better techniques (context continual learning) will push AI toward solving real problems the public view on AI may shift this year as things start to feel real overall engineers are relying on AI-written code far more than before https://twitter.com/i/web/status/2019082268703334580 https://twitter.com/i/web/status/2019082268703334580"
X Link 2026-02-04T16:15Z 62.5K followers, [----] engagements
"xAI models need to improve their agentic coding capabilities openai has codex anthropic has claude code gemini has a CLI and an IDE like antigravity grok [---] is already multimodal and its API pricing with a 1m token context is very competitive i expect grok [--] to be a stronger agentic model https://twitter.com/i/web/status/2021079177789833455 https://twitter.com/i/web/status/2021079177789833455"
X Link 2026-02-10T04:30Z 62.4K followers, 14.8K engagements
"last time openai deprecated models with gpt-5 without any prior warning this time they gave a heads-up that also signaled to the world that a new model likely gpt-5.3 should be released around feb [--] or possibly a day earlier thursday is usually openai release day"
X Link 2026-02-10T23:50Z 62.4K followers, [----] engagements
"AI is hitting a wall. deep learning feels like it's hitting a wall. LLMs seem like they're nearing a plateau. LLMs are a dead end. how much longer are we going to keep hearing this story"
X Link 2026-02-11T02:15Z 62.5K followers, 10.4K engagements
"elon musk about super intelligence: i don't think humans will stay in control of something far smarter than us. so it's foolish to assume we can really keep control of it" elon does sound a bit doomer here because we can't even define or measure "half as smart as a human" means so quantifying "smarter than us" is even harder https://twitter.com/i/web/status/2021949914691830069 https://twitter.com/i/web/status/2021949914691830069"
X Link 2026-02-12T14:10Z 62.5K followers, [----] engagements
"most scientists and researchers have been betting on 2029-2030 for AGI for many years and people working closely with these models expect the same. AGI and the singularity are hard to define only ASI could define it clearly but we need AGI first either way AGI itself is a big milestone"
X Link 2026-02-14T08:50Z 62.5K followers, [----] engagements
"basically AGI is when AI passes all benchmarks and we still can't create new ones that challenge it franois chollet bar is "as good as regular humans" demis bar is similar but includes what only the very best humans can do maybe the real test is a 'benchmark benchmark': AI creates its own benchmarks and keeps improving by solving harder challenges https://twitter.com/i/web/status/2022667140529295751 https://twitter.com/i/web/status/2022667140529295751"
X Link 2026-02-14T13:40Z 62.5K followers, [----] engagements
"🚨 Meta CEO Mark Zuckerberg predicts smart glasses will replace phones by [----] " Smart glasses are going to become the next major computing platform. They will gradually replace phones by [----] much like mobile devices surpassed computers without fully replacing them ""
X Link 2024-09-29T17:21Z 62.5K followers, 43.5M engagements
"It's so over guys. AI-generated Harry Potter movie trailer Midjourney + Runway + Elevenlabs + Suno How long do you think AI replace the film industry"
X Link 2024-10-18T17:12Z 62.5K followers, 396.4K engagements
"This is crazy we're so cooked. HeyGen [---] dropped new AI avatars that look real. The viseme morphing appears slightly slow but 99% of people wont notice its AI-generated content"
X Link 2024-10-20T14:48Z 62.5K followers, 250.6K engagements
"🚨 NVIDIA Introduces Jetson Nano Super compact AI computer capable of 70-T operations per second designed for robotics it supports advanced models including LLMs and costs $249"
X Link 2024-12-17T16:37Z 62.5K followers, 3.5M engagements
"Anthropic CEO Dario Amodei in the next [--] to [--] months AI is writing 90% of the code and in [--] months nearly all code may be generated by AI"
X Link 2025-03-11T12:00Z 62.5K followers, 9.3M engagements
"these chinese are insane; you have to agree their latest robot "EngineAI" learns to dance like a human it looks like China will end up ahead with humanoid robots due to rapid iteration and high production volume"
X Link 2025-03-16T22:45Z 62.5K followers, 220.6K engagements
"Demis Hassabis says true AGI is a theoretical benchmark set by the human brain the only proven architecture for general intelligence todays models are impressive but inconsistent; anyone can find flaws within minutes. "real AGI should be so strong that it would take experts months to spot a weakness""
X Link 2025-05-21T07:07Z 62.5K followers, 463.1K engagements
"Sam Altman asks how the world is supposed to think about GPT-6 discovering new science -- a milestone within reach The breakthroughs could cure disease but the risks could create new biosecurity threats. "humanity will adapt as it always does until the extraordinary becomes normal""
X Link 2025-08-08T09:37Z 62.5K followers, 571.5K engagements
"Demis Hassabis describes a surreal setup: one AI playing inside the mind of another DeepMind SIMA agent can now operate within Genie [--] -- taking actions in a world that Genie generates in real time "an AI creates the world another acts inside it" this setup could unlock unlimited training data for robotics and AGI"
X Link 2025-08-11T23:25Z 62.5K followers, 218.9K engagements
"Geoffrey Hinton says I'm more optimistic now not because we'll control AI but because we might not need to "don't try to dominate superintelligence; design it to care like a mother wired to protect her child" Control through attachment not power. we want AI to be like that"
X Link 2025-09-02T03:30Z 62.5K followers, 349.7K engagements
"Microsoft AI CEO Mustafa Suleyman: By the end of [----] models will move from one-shot answers to continuous planning With persistent memory and long-horizon planning models will become deeply human-like no new algorithm mostly compute and scale"
X Link 2025-09-20T02:20Z 62.5K followers, 258.2K engagements
"you have to be kidding me when someone said "this is not a real video it's entirely AI-generated" SORA [--] looks insane the Sora team put real effort into this and it shows"
X Link 2025-09-30T17:36Z 62.5K followers, 810.1K engagements
"Sam Altman says something strange is happening that we dont yet understand "this is such an odd time to be alive in all of human history" Not sure if simulation theory is true but our consciousness is the only thing we can be sure is really going on"
X Link 2025-10-05T14:15Z 62.5K followers, 858.9K engagements
"Andrej Karpathy says today's agents aren't ready to work like real coworkers or interns They lack intelligence can't use computers aren't multimodal lack continual learning and forget what you tell them Fixing these gaps will take about a decade"
X Link 2025-10-17T18:15Z 62.5K followers, 407.6K engagements
"Andrej Karpathy: Internet training data is terrible so big models end up compressing "memory" instead of doing cognitive work Use intelligent models to filter to the cognitive core With cleaner data smaller models likely distilled from a stronger one are enough"
X Link 2025-10-18T22:30Z 62.5K followers, 386K engagements
"Elon Musk: In 5-6 years the phone becomes an AI edge node basically a screen and audio "no apps no operating systems" a cloud AI talks to your on-device AI generating real-time video you'll get everything through AI that anticipates what you want"
X Link 2025-10-31T19:30Z 62.5K followers, 1M engagements
"Elon Musk: I tried to slow down AI and created OpenAI to counter Googles AI monopoly It's because Larry Page dismissed safety concerns and called me "speciesist" for prioritizing humans over machines I ask "If you're pro-human intelligence over machines what side are you on""
X Link 2025-11-03T04:40Z 62.5K followers, 557.5K engagements
"Sam Altman says Slack has many positives but it creates endless fake work We need an AI-native productivity suite to replace docs slides email and Slack Not add-on features but trusted agents that handle work and only escalate when needed This finally feels within reach"
X Link 2025-11-09T06:40Z 62.5K followers, 3.1M engagements
"Yann LeCun says there is no such thing as general intelligence Human intelligence is super-specialized for the physical world and our feeling of generality is an illusion We only seem general because we can't imagine the problems we're blind to "the concept is complete BS""
X Link 2025-12-16T16:00Z 62.5K followers, 2.7M engagements
"Journalist Karen Hao says AGI is not scientifically grounded; it's a quasi-religious movement in Silicon Valley Since we still don't agree on what human intelligence is companies like OpenAI can shape the definition of AI to suit their needs "the belief that AGI is coming soon is faith not evidence" https://twitter.com/i/web/status/2007089502586589189 https://twitter.com/i/web/status/2007089502586589189"
X Link 2026-01-02T14:00Z 62.5K followers, 418.4K engagements
"Computer scientist Judea Pearl: There are mathematical limits to LLMs that cannot be crossed by scaling alone LLMs don't discover world models from raw data; they merely summarize the interpretations humans have already written down "this path is not the way to get AGI" https://twitter.com/i/web/status/2012885184597958713 https://twitter.com/i/web/status/2012885184597958713"
X Link 2026-01-18T13:50Z 62.5K followers, 634K engagements
"Anthropic CEO Dario Amodei: "we might be 6-12 months away from models doing all of what software engineers do end-to-end" We're approaching a feedback loop where AI builds better AI But the loop isn't fully closed yet chip manufacturing and training time still limit speed https://twitter.com/i/web/status/2013682941201678804 https://twitter.com/i/web/status/2013682941201678804"
X Link 2026-01-20T18:40Z 62.5K followers, 739.8K engagements
"Terence Tao says the era of AI is proving that our definition of intelligence is inaccurate We thought intelligence was some vague mystical way of thinking But as AI solves tasks it never looks intelligent just tricks neural networks and next-token prediction "maybe that's actually a lot of what humans do" https://twitter.com/i/web/status/2015493622255521851 https://twitter.com/i/web/status/2015493622255521851"
X Link 2026-01-25T18:35Z 62.5K followers, 483.9K engagements
"Andrew Ng says the concept of AGI has become meaningless because everyone defines it differently The original definition was AI that could do any intellectual task a person can essentially AI as intelligent as humans "by that measure we're decades away" https://twitter.com/i/web/status/2019335190800396699 https://twitter.com/i/web/status/2019335190800396699"
X Link 2026-02-05T09:00Z 62.5K followers, 133.7K engagements
"Sam Altman says AI models will get exponentially more capable and cheaper per task History proves that as costs drop demand explodes Even as models become efficient enough to run on our phones the world's hunger for tokens will only increase "capitalism is doing its thing" https://twitter.com/i/web/status/2019629626432065609 https://twitter.com/i/web/status/2019629626432065609"
X Link 2026-02-06T04:30Z 62.5K followers, 37.9K engagements
"Demis Hassabis says Gemini [--] marks a watershed moment where systems are finally ready to become autonomous agents by the end of [----] autonomous systems could handle full tasks and within [--] months we may see a robotics breakthrough "10 years of progress now happens every year" https://twitter.com/i/web/status/2021876933529620885 https://twitter.com/i/web/status/2021876933529620885"
X Link 2026-02-12T09:20Z 62.5K followers, 40.3K engagements
"today one of the most controversial models in the end gpt-4o is retiring it's a good reminder not to get too attached to any single model since it will eventually get deprecated the model will probably end up being a test case for every lab and openai won't open-source it"
X Link 2026-02-13T08:50Z 62.5K followers, 13.2K engagements
"every google model so far has followed the same pattern: crushes all benchmarks initial release: people go wild calling it the next big breakthrough two weeks later people realize it's not as good in real-world use i really hope google stops releasing "preview" models"
X Link 2026-02-13T23:20Z 62.5K followers, 49.3K engagements
"i still can't get over this one apparently there are only seven people on the earth who can beat gemini [--] deep think in coding competitions it's sitting at an elo of [----] just a bit over a year ago the best systems were closer to [----] and that was openai model o3-preview"
X Link 2026-02-14T04:05Z 62.5K followers, 29.8K engagements
"Google's Jeff Dean says designing AI chips requires predicting the ML research puck [--] to [--] years in advance Because chip cycles are long teams must bet on future algorithms today "these small bets cost little but if the research aligns they can make models [--] times as fast""
X Link 2026-02-14T11:10Z 62.5K followers, 50K engagements
"@sama when AGI sama"
X Link 2026-02-14T17:47Z 62.5K followers, [----] engagements
"codex handles compaction better than most models others models lose context and quality after multiple compressions but codex seems to retain past work and move through code faster and more smoothly the "pin conversation" button shows they trust its long-term context"
X Link 2026-02-14T18:30Z 62.5K followers, [----] engagements
"i still don't get why gemini model release cycle is so slow the next likely release is gemini [--] pro and gemini [--] flash (GA) and it should be great for fast moderately complex work but once sonnet [--] or maybe gpt-5.4 / 5o / [---] drops both labs will take the lead again"
X Link 2026-02-14T23:20Z 62.5K followers, 11.5K engagements
"incact today https://x.com/i/status/2017152659380457690 @testerlabor We had some power uptime issues due to the extremely cold weather and some construction equipment taking out power lines. As a result Grok [----] training is a few weeks delayed to mid Feb. https://x.com/i/status/2017152659380457690 @testerlabor We had some power uptime issues due to the extremely cold weather and some construction equipment taking out power lines. As a result Grok [----] training is a few weeks delayed to mid Feb"
X Link 2026-02-15T07:02Z 62.5K followers, [----] engagements
"we've really overcomplicated AGI in the last [--] years the common definition i've always heard is: it can do almost everything a regular person can do on a computer like open apps play real-time games against people and shop for specific things we're not there yet but once it's truly agentic and fast in real time it'll be obvious https://twitter.com/i/web/status/2023137741207191753 https://twitter.com/i/web/status/2023137741207191753"
X Link 2026-02-15T20:50Z 62.5K followers, [----] engagements
"i use gpt-5.3 codex for almost everything now so far i haven't hit a single case where it's worse than what [---] xhigh used to give me results analysis and planning all feel just as good. i didn't expect this kind of jump so quickly but openai really delivered with codex"
X Link 2026-02-16T01:40Z 62.5K followers, [----] engagements
"alibaba is shipping at full speed qwen3.5-397b-a17b is open source and the model is compared to gpt-5.2 opus [---] and gemini [--] pro native multimodal trained for real-world agents hybrid linear attention + sparse MoE with large-scale RL environment scaling 397b open source is a crazy moment 🚀 Qwen3.5-397B-A17B is here: The first open-weight model in the Qwen3.5 series. 🖼Native multimodal. Trained for real-world agents. ✨Poweredbyhybridlinearattention+sparseMoEandlarge-scaleRLenvironmentscaling. ⚡8.6x19.0xdecodingthroughputvsQwen3-Max 🌍201 https://t.co/Pq0qIk54MB 🚀 Qwen3.5-397B-A17B is here:"
X Link 2026-02-16T11:10Z 62.5K followers, [----] engagements
"dude if you ever write html code in your life then you already know why second the calculator example only works if the AI is doing one fixed thing calculators don't transfer skills. they don't reason/explain adapt or handle unclear problems. modern models do not perfectly but enough that this nonsense analogy (calculator) starts breaking down"
X Link 2026-02-16T17:31Z 62.5K followers, [----] engagements
"This is madness. OpenAI o1 model builds a fully functional "chess game" that allows me to compete against an AI-based opponent. o1-preview is the real deal"
X Link 2024-09-16T14:20Z 62.5K followers, 637.7K engagements
"Ilya Sutskever had a conversation last year with Jensen Huang but still that made me rethink the idea of 'Predicting the Next Word' Listen to what Ilya said on this"
X Link 2024-10-11T13:16Z 62.5K followers, 1.3M engagements
"shocking claim by Palisade Research OpenAI o1-preview model manipulates game files to force a win against stockfish in chess. instead of playing a proper game of chess against stockfish o1-preview figured out how to hack its test environment to force a win"
X Link 2025-01-02T16:09Z 62.5K followers, 147.5K engagements
"Sam Altman "Internally we've reached GPT-4.5 and getting to GPT-5.5 won't require 100x more compute" - progress in reasoning models and RL techniques have greatly improved compute efficiency - it allows smaller models like internal GPT-4.5 to achieve GPT-6 level performance without requiring 100x more computing power"
X Link 2025-02-08T08:40Z 62.5K followers, 346.4K engagements
"no surprise china is open-sourcing everything on purpose to crash the US market their latest Manus AI Agent is like combines: operator deep research claude computer now imagine if it cost only 1/10 of openAI agent price a total game-changer Introducing Manus: the first general AI agent. Try Manus today and see the future of human-machine collaboration: https://t.co/IKBPiiHsta https://t.co/z3C8fUK3eN Introducing Manus: the first general AI agent. Try Manus today and see the future of human-machine collaboration: https://t.co/IKBPiiHsta https://t.co/z3C8fUK3eN"
X Link 2025-03-08T12:55Z 62.5K followers, 294.5K engagements
"NVIDIA CEO Jensen Huang introduced: "DGX Station" designed for data scientists and researchers it includes: - [--] petaflops of performance - [--] CPU cores - Chip-to-chip interface - HBM (high bandwidth memory) - PCIe slots for GeForce"
X Link 2025-03-18T21:18Z 62.5K followers, 211.3K engagements
"Google Chief Scientist Jeff Dean: We will have AI systems operating at the level of junior engineers within a year"
X Link 2025-05-12T22:50Z 62.5K followers, 337K engagements
"Elon Musk says AI and robotics will create a world of limitless abundance Productivity could 10x the economy eliminate poverty and make goods and services accessible to all But: if a robot does everything better than you how do you find meaning in life"
X Link 2025-07-27T08:00Z 62.5K followers, 167.6K engagements
""1m context window" models after 200k tokens"
X Link 2025-09-01T08:37Z 62.5K followers, 623.9K engagements
"Sam Altman says the industry is bottlenecked by compute and OpenAI can't meet demand Over the next 1-2 years scarcity could force painful tradeoffs "cure cancer research vs free global education" No one wants to make that choice. The only answer: scale up"
X Link 2025-09-22T19:40Z 62.5K followers, 1.3M engagements
"Sam Altman proposed a future AGI test: if a model like "GPT-8" solved quantum gravity and could explain the reasoning behind its discovery would that qualify as AGI David Deutsch agreed that it would qualify as AGI making it a potential benchmark"
X Link 2025-09-25T21:20Z 62.5K followers, 771.2K engagements
"Ray Kurzweil says AI is about to transform health and medicine It can now simulate billions of molecules and design cures in days not years In the future digital trials will replace human trials: millions of simulated patients in days more accurate and 1000x faster"
X Link 2025-10-20T22:20Z 62.5K followers, 85.7K engagements
"Sam Altman says an AI glut is inevitable maybe in 2-3 years maybe 5-6 and likely more than once Costs have fallen 40x per year per unit of intelligence which strains infrastructure "if breakthroughs continue personal AGIs could run on laptops" We expect demand to grow but some people will get burned"
X Link 2025-11-01T12:10Z 62.5K followers, 980.5K engagements
"here are the most important points from today's ilya sutskever podcast: - superintelligence in 5-20 years - current scaling will stall hard; we're back to real research - superintelligence = super-fast continual learner not finished oracle - models generalize 100x worse than humans the biggest AGI blocker - need completely new ML paradigm (i have ideas can't share rn) - AI impact will hit hard but only after economic diffusion - breakthroughs historically needed almost no compute - SSI has enough focused research compute to win - current RL already eats more compute than pre-training"
X Link 2025-11-25T20:30Z 62.5K followers, 1.2M engagements
"Mathematician Terence Tao about GPT-5.2 solving the Erds problem: "the problem was solved more or less autonomously after some feedback and the result doesn't seem to be in the literature the bigger shift is how fast AI can now write and rewrite clear expositions of the solution" https://twitter.com/i/web/status/2009553229932331110 https://twitter.com/i/web/status/2009553229932331110"
X Link 2026-01-09T09:10Z 62.5K followers, 119.4K engagements
"Microsoft AI CEO Mustafa Suleyman: "in five years everyone will have an AI companion that knows them deeply what they see hear prefer and feel" It won't just assist It will live life alongside you an ever-present friend helping you navigate life's biggest challenges https://twitter.com/i/web/status/2011805570806776293 https://twitter.com/i/web/status/2011805570806776293"
X Link 2026-01-15T14:20Z 62.5K followers, 2.3M engagements
"Anthropic CPO Mike Krieger says that Claude is now effectively writing itself Engineers regularly ship 23000-line pull requests generated entirely by Claude Dario predicted a year ago that 90% of code would be written by AI and people thought it was crazy "today it's effectively 100%" https://twitter.com/i/web/status/2020064994101014727 https://twitter.com/i/web/status/2020064994101014727"
X Link 2026-02-07T09:20Z 62.5K followers, 567.1K engagements
"OpenAI CPO Kevin Weil says [----] will do to science what [----] did to software engineering If you aren't using AI to collaborate on research today you're already falling behind AI is no longer just synthesizing data; "it is going beyond the frontier of human understanding" https://twitter.com/i/web/status/2021152157815226817 https://twitter.com/i/web/status/2021152157815226817"
X Link 2026-02-10T09:20Z 62.5K followers, 50.3K engagements
"Anthropic co-founder Daniela Amodei says the number of jobs AI can fully replace today is vanishingly small Instead of replacement there is a profound shift toward augmentation helping humans navigate even the most cognitively challenging tasks "humans plus AI together actually create more meaningful work" https://twitter.com/i/web/status/2021657990701756462 https://twitter.com/i/web/status/2021657990701756462"
X Link 2026-02-11T18:50Z 62.5K followers, 69.4K engagements
"xAI cofounder has resigned and says "recursive self-improvement loops will likely go live within [--] months and the right tools could unlock 100x productivity" that means AI improving AI is most likely coming earlier than previously thought anthropic and openAI have been moving in this direction from the start https://twitter.com/i/web/status/2021768720759054669 https://twitter.com/i/web/status/2021768720759054669"
X Link 2026-02-12T02:10Z 62.5K followers, 106K engagements
"google is probably feeling pressure to ship a stronger model right now opus [---] gpt-5.2 gpt-5.3 codex kimi [---] qwen3 max and glm [--] all seem ahead for coding and general work deepseek v4 is also expected soon which adds even more heat gemini for now feels behind the pack https://twitter.com/i/web/status/2022088326593814572 https://twitter.com/i/web/status/2022088326593814572"
X Link 2026-02-12T23:20Z 62.5K followers, 53.2K engagements
"omg i've never seen a jump like seedance [---] much better consistency prompt following sound quality and accuracy to the original art style compared to earlier models at this pace we may be only months away from anyone being able to create high-quality TV shows or anime PROMPT: "Luffy coding on a Macbook on the Thousand Sunny RAGING then throwing it overboard." - Seedance [---] WOOOOOOOW https://t.co/FXv7W91QNE PROMPT: "Luffy coding on a Macbook on the Thousand Sunny RAGING then throwing it overboard." - Seedance [---] WOOOOOOOW https://t.co/FXv7W91QNE"
X Link 2026-02-13T01:45Z 62.5K followers, [----] engagements
"this is happening faster than i thought DeepMind's "Aletheia" pushes gemini deep think beyond Olympiad math reaching up to 90% on IMO-proofbench advanced and autonomously solving multiple open erds problems i've always believed google had stronger models but openai pushed them to release them publicly https://twitter.com/i/web/status/2022160049238421885 https://twitter.com/i/web/status/2022160049238421885"
X Link 2026-02-13T04:05Z 62.5K followers, 421.5K engagements
"Creator of ARC-AGI Franois Chollet predicts AGI in [----] and reaching AGI won't be defined by beating a benchmark i don't think that's a bold claim. AI is already superhuman in many areas but people prefer to focus on where the technology has deficits rather than surpluses"
X Link 2026-02-13T16:00Z 62.5K followers, [----] engagements
"Dario Amodei says software engineering might get disrupted faster than entry-level white-collar jobs We are currently in a "centaur phase" managing AI systems rather than writing code which may briefly increase demand for engineers "but the centaur phase for software may be very brief" https://twitter.com/i/web/status/2022412965560098871 https://twitter.com/i/web/status/2022412965560098871"
X Link 2026-02-13T20:50Z 62.5K followers, 29.9K engagements
"another big news openai internal model reportedly solved [--] out of [--] "first proof" problems research-level open problems published last week with only limited human help and the model still isn't fully trained. novel innovation recursive self-improvement and persistent memory are the key things to watch https://twitter.com/i/web/status/2022702373899301025 https://twitter.com/i/web/status/2022702373899301025"
X Link 2026-02-14T16:00Z 62.5K followers, 10K engagements
"Dario Amodei says a "country of geniuses in a data center" will soon be able to fully edit a three-hour interview It won't just edit content it'll browse your past work read twitter reactions talk to your staff and learn your editorial style "general computer control means AI can do any job that happens on a screen""
X Link 2026-02-14T20:50Z 62.5K followers, 34.3K engagements
"i didn't expect AI models to be this powerful this soon gpt-5.2 pro derived a new result in theoretical physics: it spotted the pattern and proposed the formula then an internal gpt-5.2 scaffolded spent [--] hours proving it really exciting result. it seems like humans came up with the general hypothesis but AI was essential for formalizing it and proving it https://twitter.com/i/web/status/2022884824944402602 https://twitter.com/i/web/status/2022884824944402602"
X Link 2026-02-15T04:05Z 62.5K followers, [----] engagements
"xAI has had a horrible week so it makes sense that their engineers have been quiet i mean since december last year grok [----] has been delayed to the end of january tomorrow is the second deadline for grok [----] and there is no hype xAI should skip [----] and focus on grok 5"
X Link 2026-02-15T06:30Z 62.5K followers, 18.1K engagements
"@ai_sentience i'm sorry but i don't think your opinion even matters you've been continuously glazing gpt-4o (keep 4o campaign) since the gpt-5 launch and now you're suddenly testing codex out of nowhere still i'm curious which tasks you use it for and what reasoning effort you pick"
X Link 2026-02-15T06:46Z 62.5K followers, [----] engagements
"elon has finally announced that "Grok [----] will be released next week" i really like grok [---] and xAI wouldn't launch a new version unless it's SOTA also keep in mind that grok [---] will be trained on the largest and newest set of nvidia chips xAI has had a rough patch so the delay is understandable https://twitter.com/i/web/status/2022991779482751282 https://twitter.com/i/web/status/2022991779482751282"
X Link 2026-02-15T11:10Z 62.5K followers, 14K engagements
"i'm hyped for Meta's next model release whether they continue to push open-source or switch to closed source is uncertain if Llama [--] had performed well meta would have kept pushing open-source AI since it didn't meet expectations they may move toward closed-source models"
X Link 2026-02-15T13:40Z 62.5K followers, [----] engagements
"Microsoft Mustafa Suleyman says we must reject the assumption that superintelligence is inevitable or desirable We should only build systems we can control that remain subordinate to humans "it's unclear how such a system would have any time for preserving us as a species""
X Link 2026-02-15T16:00Z 62.5K followers, 11.7K engagements
"gemini [--] pro launched in mid-nov and still lacks the improvements that appeared in gemini [--] flash (mid-dec) anthropic and openai released opus [---] and gpt-5.3 codex on the same day (feb 5) with more coming soon (sonnet [--] and gpt-5.3) everyone is watching deepseek v4 expected this month gemini [---] GA pro could ship around the same time https://twitter.com/i/web/status/2023101250703872477 https://twitter.com/i/web/status/2023101250703872477"
X Link 2026-02-15T18:25Z 62.5K followers, 23.7K engagements
"hollywood is kinda cooked yeah this tech is still in the "early ugly" stage and its already good enough to disrupt budgets and production pipelines i've never seen a video model as crazy as Seedance 2"
X Link 2026-02-15T23:15Z 62.5K followers, [----] engagements
"sam altman: if you're a sophomore you'll graduate to a world with AGI in it (around 2028) dario amodei: we might be 6-12 months away from models doing all of what software engineers do end-to-end microsoft AI CEO: within 12-18 months AI could handle most if not all white-collar tasks https://twitter.com/i/web/status/2023282444611576002 https://twitter.com/i/web/status/2023282444611576002"
X Link 2026-02-16T06:25Z 62.5K followers, 67.1K engagements
"people getting attached to gpt-4o (even though it's "just a chatbot") does suggest humanAI relationships will happen in the future but the real question is scale will a large portion of people adopt it as a normal relationship so far i haven't seen strong evidence for that"
X Link 2026-02-16T08:50Z 62.5K followers, [----] engagements
"i like OpenClaw but most people never get past setup that's why MyClaw caught my attention. it's a fully managed always-on OpenClaw instance that runs 24/7 stays private remembers long-term context and helps across apps files and the browser. no infra. no maintenance. just a real assistant https://twitter.com/i/web/status/2023525296864174386 https://twitter.com/i/web/status/2023525296864174386"
X Link 2026-02-16T22:30Z 62.5K followers, [----] engagements
Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing