# ![@dieaud91 Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::1726570086960365568.png) @dieaud91 Diego Aud

Diego Aud posts on X about ai, open ai, human, agi the most. They currently have [---] followers and [---] posts still getting attention that total [---] engagements in the last [--] hours.

### Engagements: [---] [#](/creator/twitter::1726570086960365568/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1726570086960365568/c:line/m:interactions.svg)

- [--] Week [------] +533%
- [--] Month [-------] +1,252%
- [--] Months [-------] +114%
- [--] Year [-------] -85%

### Mentions: [--] [#](/creator/twitter::1726570086960365568/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1726570086960365568/c:line/m:posts_active.svg)

- [--] Week [--] +56%
- [--] Month [--] +163%
- [--] Months [---] +93%
- [--] Year [---] -28%

### Followers: [---] [#](/creator/twitter::1726570086960365568/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1726570086960365568/c:line/m:followers.svg)

- [--] Week [---] +0.79%
- [--] Month [---] +26%
- [--] Months [---] +53%
- [--] Year [---] +168%

### CreatorRank: [---------] [#](/creator/twitter::1726570086960365568/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1726570086960365568/c:line/m:influencer_rank.svg)

### Social Influence

**Social category influence**
[currencies](/list/currencies)  58.02% [technology brands](/list/technology-brands)  19.08% [stocks](/list/stocks)  3.82% [finance](/list/finance)  3.05% [countries](/list/countries)  0.76% [social networks](/list/social-networks)  0.76%

**Social topic influence**
[ai](/topic/ai) 27.48%, [open ai](/topic/open-ai) #2554, [human](/topic/human) 6.87%, [agi](/topic/agi) #830, [in the](/topic/in-the) 6.11%, [agents](/topic/agents) 6.11%, [this is](/topic/this-is) 6.11%, [we are](/topic/we-are) 4.58%, [model](/topic/model) 4.58%, [science](/topic/science) 3.82%

**Top accounts mentioned or mentioned by**
[@emollick](/creator/undefined) [@kimmonismus](/creator/undefined) [@openai](/creator/undefined) [@yacinemtb](/creator/undefined) [@antonpme](/creator/undefined) [@drbeavisai](/creator/undefined) [@btibor91](/creator/undefined) [@polynoamial](/creator/undefined) [@slowdeveloper](/creator/undefined) [@tsarnick](/creator/undefined) [@sama](/creator/undefined) [@artuskg](/creator/undefined) [@sciguy](/creator/undefined) [@educatingwithai](/creator/undefined) [@markk](/creator/undefined) [@keridwencodet](/creator/undefined) [@tszzl](/creator/undefined) [@applesjimmy](/creator/undefined) [@therealadamg](/creator/undefined) [@dougchampion](/creator/undefined)

**Top assets mentioned**
[Alphabet Inc Class A (GOOGL)](/topic/$googl)
### Top Social Posts
Top posts by engagements in the last [--] hours

"As I often say I don't endorse one company over another but in this case I understand OpenAI. How can people expect to get powerful AI for totally free Who pays for that ton of compute I think that if you know the basics of how the economy works you can accept that Free accounts might come with (responsible) ads. I agree that it's annoying and I certainly hope they'll be implemented in the right way but someone has to pay the bill. 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"  
[X Link](https://x.com/dieaud91/status/2019141409891041651)  2026-02-04T20:10Z [---] followers, [---] engagements


"Exciting but expected at the same time. If you use the model extensively especially Heavy Thinking or [---] Pro you know that it is the most reliable and honestly useful for real work. GPT-5.3 will be disruptive. We estimate that GPT-5.2 with high (not xhigh) reasoning effort has a 50%-time-horizon of around [---] hrs (95% CI of [--] hr [--] min to [--] hr [--] min) on our expanded suite of software tasks. This is the highest estimate for a time horizon measurement we have reported to date. https://t.co/USkHNuFexc We estimate that GPT-5.2 with high (not xhigh) reasoning effort has a 50%-time-horizon of"  
[X Link](https://x.com/dieaud91/status/2019173040412901517)  2026-02-04T22:15Z [---] followers, [---] engagements


"Things are heating up. We got to High cybersecurity. I'm drooling over GPT-5.3 Pro in ChatGPT hope we get it soon. Meanwhile I think I'll take a week to try Opus [---]. If benchmarks tell the right story it seems like a very strong model especially for Office work. This is our first model that hits "high" for cybersecurity on our preparedness framework. We are piloting a Trusted Access framework and committing $10 million in API credits to accelerate cyber defense. https://t.co/vN8KpIuzjt This is our first model that hits "high" for cybersecurity on our preparedness framework. We are piloting a"  
[X Link](https://x.com/dieaud91/status/2019478502437736891)  2026-02-05T18:29Z [---] followers, [---] engagements


"@TheRealAdamG @DrBeavisAI @adonis_singh @OpenAI @ChatGPTapp @TheRealAdamG Are GPT-5.3-Codex and the non-Codex GPT-5.3 (Thinking/Pro) identical in terms of performance If so thats disappointing.OpenAI reports GDPval is unchanged vs GPT-5.2"  
[X Link](https://x.com/dieaud91/status/2019669086247219267)  2026-02-06T07:06Z [---] followers, [--] engagements


"We need big models back. If this is the case and Claude Opus is a smaller model than Opus [---] then we are seriously underestimating the real possible frontier. The headline is Opus [---] scores 69% for $3.50/task on ARC v2. This up +30pp from Opus [---]. We attribute performance to the new "max" mode and 2X reasoning token budget -- notably task cost is held steady. Based on early field reports and other benchmark scores like SWE Bench The headline is Opus [---] scores 69% for $3.50/task on ARC v2. This up +30pp from Opus [---]. We attribute performance to the new "max" mode and 2X reasoning token"  
[X Link](https://x.com/dieaud91/status/2019674846570135592)  2026-02-06T07:29Z [---] followers, [---] engagements


"I believe the debate should shift from "When AGI" to "When AHCI" (Artificial Human-Competent Intelligence). Why If an AI model can reliably perform a huge chunk of economically valuable human tasks the implications for civilization are massive even if it doesn't solve fusion or show genuine continual learning. The AGI debate has become meaningless cos nobody can agree on the definition. If you set the target define the target youd reach it every time. We dont even have a peer reviewed study to define our intelligence & consciousness. So a more meaningful discussion should focus on The AGI"  
[X Link](https://x.com/dieaud91/status/2020472490187686110)  2026-02-08T12:19Z [---] followers, [---] engagements


"@antonpme I think that to explore the computer use part I need to set up a 'sandbox environment' where I don't have patient files 🙁 otherwise I'd be risking too much"  
[X Link](https://x.com/dieaud91/status/2020474194593530068)  2026-02-08T12:26Z [---] followers, [--] engagements


"Fuel the bio acceleration. AI for science is one of the domains I'm most excited about. We need to deploy the most powerful models in all the research pipelines and drive fast through the "cut off" for Longevity Escape Velocity The drug design engine were building at @IsomorphicLabs is extending the SOTA further across key benchmarks showing huge progress in accuracy and capabilities critical for in-silico drug discovery. Incredible work from @maxjaderberg and the entire team at Isomorphic Labs The drug design engine were building at @IsomorphicLabs is extending the SOTA further across key"  
[X Link](https://x.com/dieaud91/status/2021231246903287958)  2026-02-10T14:34Z [---] followers, [---] engagements


"@artuskg Thanks for the tag. As a clinical nutritionist reliability is everything. It's reassuring to see research focusing on calculation accuracy rather than just memorization. We need tools we can trust with patient data. Great work on the audit"  
[X Link](https://x.com/dieaud91/status/2021264585811013664)  2026-02-10T16:46Z [---] followers, [--] engagements


"Nice update but why go for [---] if [---] is right around the corner Or is it. 💔 Deep research in ChatGPT is now powered by GPT-5.2. Rolling out starting today with more improvements. https://t.co/LdgoWlucuE Deep research in ChatGPT is now powered by GPT-5.2. Rolling out starting today with more improvements. https://t.co/LdgoWlucuE"  
[X Link](https://x.com/dieaud91/status/2021306223417561123)  2026-02-10T19:32Z [---] followers, 19.3K engagements


"@sciguy Fair point regarding stability I get that. But since Codex [---] is already out and not everyone is a coder itd be quite disappointing if the standard [---] got delayed for weeks"  
[X Link](https://x.com/dieaud91/status/2021308134573113495)  2026-02-10T19:39Z [---] followers, [----] engagements


"@sciguy Fingers crossed 5.3-Pro can't come soon enough"  
[X Link](https://x.com/dieaud91/status/2021313930044313847)  2026-02-10T20:02Z [---] followers, [---] engagements


"@DougChampion Could be. Honestly I just hope we'll get "general" [---] as soon as possible. If it's smarter AND faster I can already see the huge benefits for automating my workflows"  
[X Link](https://x.com/dieaud91/status/2021354083031159035)  2026-02-10T22:42Z [---] followers, [---] engagements


"@DrBeavisAI @btibor91 If the model we get is just a somewhat better [---] and we don't get a boost in key metrics besides conversation I think I'll give Opus [---] a try and buy a Claude sub for [--] month. [---] is good but something tells me Opus [---] might be a little stronger"  
[X Link](https://x.com/dieaud91/status/2021379403171971535)  2026-02-11T00:22Z [---] followers, [---] engagements


"For me the only good scenario in this situation is if they deploy [---] Thinking and Pro skipping Instant for some time. It would be extremely disappointing to see [---] inside Codex (atm way over my head) and knowing it is many weeks away in ChatGPT. We updated GPT-5.2 (the instant model) in ChatGPT today. Not a huge change but hopefully you find it a little better. We updated GPT-5.2 (the instant model) in ChatGPT today. Not a huge change but hopefully you find it a little better"  
[X Link](https://x.com/dieaud91/status/2021485596146028949)  2026-02-11T07:24Z [---] followers, [----] engagements


"@EducatingwithAI Yep about 60% of my work relies on MS Office too. That's why [---] Pro was the first really useful model and that's also why I'm eager to get [---] (hoping for a boost in exactly this.)"  
[X Link](https://x.com/dieaud91/status/2021567671012147456)  2026-02-11T12:51Z [---] followers, [--] engagements


"@EducatingwithAI If [---] gets delayed I'm 100% trying a month of Claude with Opus 4.6"  
[X Link](https://x.com/dieaud91/status/2021580534745280739)  2026-02-11T13:42Z [---] followers, [--] engagements


"The new Google Champion is coming. Gemini [---] is on the horizon say the line @OfficialLoganK say the line @OfficialLoganK"  
[X Link](https://x.com/dieaud91/status/2021715779389911135)  2026-02-11T22:39Z [---] followers, [----] engagements


"My gut tells me this week we are NOT getting [---] in ChatGPT but something different. I doubt it'll excite me more than [---] Pro (faster) would have"  
[X Link](https://x.com/dieaud91/status/2021929553808044400)  2026-02-12T12:49Z [---] followers, [----] engagements


"Exactly. Something doesn't add up. Why bother with a journal article about a new chat model if it's just a small upgrade (quoting Sam: "hopefully you find It a little better") of the same model Is this just a consolation prize for GPT-4o users I don't know. This "first Codex then one day ChatGPT" vibe feels weird https://twitter.com/i/web/status/2021933036271464918 https://twitter.com/i/web/status/2021933036271464918"  
[X Link](https://x.com/dieaud91/status/2021933036271464918)  2026-02-12T13:02Z [---] followers, [---] engagements


"Maybe I'm just "behind" but pushing coding-focused models gated behind Codex (app/CLI/IDE extensions) is not the way for mass adoption (maybe it's intentional to release it "gradually"). As a "non-technical" person new to coding and IDEs and with so much on my plate in my profession I dont have time to learn how to use Codex + an IDE workflow. Im wondering if this is the direction and the chat interface is being slowly abandoned. Generally speaking adapting to new interfaces is not a problem we can learn - but right now using [---] inside Codex is too high a barrier for me and I suspect for"  
[X Link](https://x.com/dieaud91/status/2022022282894262290)  2026-02-12T18:57Z [---] followers, [---] engagements


"@artuskg Thanks I'll DM you. To be honest I'm still a bit unsure if it fits my workflow since I don't really rely on coding. 😕 but I appreciate the help anyway"  
[X Link](https://x.com/dieaud91/status/2022026612955656631)  2026-02-12T19:14Z [---] followers, [--] engagements


"@vaultminds Well if it's "just" [---] Codex Spark and nothing else it seems their post was mostly directed to developers/technical users. Nice but nothing to see for me"  
[X Link](https://x.com/dieaud91/status/2022027440336580973)  2026-02-12T19:18Z [---] followers, [--] engagements


"@CristianXIV I guess it depends. For my use cases is a nothingburger but I guess developers/technical users will be in awe"  
[X Link](https://x.com/dieaud91/status/2022027803533930746)  2026-02-12T19:19Z [---] followers, [--] engagements


"@shaunralston Something tells me we'll have a period without a new general omni model"  
[X Link](https://x.com/dieaud91/status/2022038821504958877)  2026-02-12T20:03Z [---] followers, [--] engagements


"@antonpme Indeed totally with you on this. China will set the pace of acceleration thanks to looser regulations strong determination + a massive talent pool and frankly I'm all for it 🔥"  
[X Link](https://x.com/dieaud91/status/2022088552788758805)  2026-02-12T23:20Z [---] followers, [--] engagements


"@polynoamial @fchollet Curious about your timeline (as an insider) for an AI that can automate most economically valuable human tasks. How long do you see it I suspect this economic milestone will hit much sooner than true generalized "AGI""  
[X Link](https://x.com/dieaud91/status/2022090918724317357)  2026-02-12T23:30Z [---] followers, [----] engagements


"@simone_m_romeo Thanks honestly I'm a complete newbie. The issue is the setup: repositories Github. It's all new to me. When I open Codex I feel lost. It feels like a steep learning curve and I simply don't have the time to climb it right now with clients breathing down my neck 🥲"  
[X Link](https://x.com/dieaud91/status/2022212020565954878)  2026-02-13T07:31Z [---] followers, [--] engagements


"My main grudge right now is that GPT-5.3 Pro exists somewhere in the world constrained in the chat of a few that are silently red teaming it and I can't access it until who knows when @aidan_mclau Wen 5.3-pro-spark @aidan_mclau Wen 5.3-pro-spark"  
[X Link](https://x.com/dieaud91/status/2022221665929978137)  2026-02-13T08:09Z [---] followers, [----] engagements


"@antonpme @mark_k Agreed. It's 'spiky' but I still find it useful for writing and creative tasks. Perhaps that spikiness is exactly what I like: sometimes its behaviors and insights make it feel. human. Can't wait for the next iteration"  
[X Link](https://x.com/dieaud91/status/2022279050396361041)  2026-02-13T11:57Z [---] followers, [--] engagements


"@KeridwenCodet @OpenAI Try Gemini [--] Pro. In my experience while a bit "quirky" it's quite helpful with translations/writing corrections"  
[X Link](https://x.com/dieaud91/status/2022452113503338810)  2026-02-13T23:25Z [---] followers, [--] engagements


"@KeridwenCodet @OpenAI Yeah for real sometimes it stirs up a bit of drama 🤣 but it still delivers. I was a heavy Claude user in [----] I think I need to give it another go even if my wallet is crying haha"  
[X Link](https://x.com/dieaud91/status/2022580704656179360)  2026-02-14T07:56Z [---] followers, [--] engagements


"@TFWNicholson Yeah the new Deep Think seems strong from what I read. Still I think they need to bake in GPT's stronger search and tool calling capabilities"  
[X Link](https://x.com/dieaud91/status/2022373890819870870)  2026-02-13T18:14Z [---] followers, [--] engagements


"This 10000x acceleration vs Moore's Law isn't just about compute power. It's about entering a world where: - Models evolve in weeks not years - Our understanding of AI gets outdated monthly - The line between possible/impossible blurs constantly Huang talks about the next decade but I'm willing to bet that even the next [--] years will feel VERY different from the last [--]. #AI #AGI #LLMs #ScalingLaws Jensen Huang says AI Scaling Laws will increase computing power by a million times in the next decade compared to Moore's Law which increased at [---] times per decade https://t.co/4bJsGV0mfK Jensen"  
[X Link](https://x.com/anyuser/status/1858600261272703017)  2024-11-18T19:56Z [---] followers, [---] engagements


"@emollick This sounds like the start of an exponential: something that will redefine the pace of academia which has notoriously always been slow"  
[X Link](https://x.com/anyuser/status/1870562899468689443)  2024-12-21T20:12Z [---] followers, [----] engagements


"@slow_developer Building in the dark in silence for the whole year"  
[X Link](https://x.com/anyuser/status/1874112312858337608)  2024-12-31T15:16Z [---] followers, [---] engagements


"The "concept" is right but they still need to create agents that are reliable enough smart enough and "capable" enough. We need them to perform meaningful chains of actions not just a few trivial tasks. When agents can reliably execute smart and complex chains of actions in Office that's when everything will change"  
[X Link](https://x.com/anyuser/status/1877454788058587391)  2025-01-09T20:37Z [---] followers, [---] engagements


"@mark_k @GoogleDeepMind @OpenAI My gut says we won't get [---]. The update to [---] was a huge letdown and doesn't point to an upcoming [---]. Hope other labs release models more powerful than [---] so that OpenAI is 'forced' to release [---] (and hopefully [---] Pro)"  
[X Link](https://x.com/dieaud91/status/2022999671275487295)  2026-02-15T11:41Z [---] followers, [---] engagements


"@BorisMPower Incredibly disappointing if this means GPT-5 has just started training and we the paying users get nothing impactful for many months if not a year. GPT-4o is good but. Not quite "GPT-4.5ish""  
[X Link](https://x.com/anyuser/status/1795478857044066805)  2024-05-28T15:34Z [---] followers, [----] engagements


"@kimmonismus I'm starting to feel that there's just nothing to be released. Patience Jimmy = wait for November when we'll show you the live DEMO of GPT-5 that will be rolled out in the "coming weeks". (TM)"  
[X Link](https://x.com/anyuser/status/1800440889086783882)  2024-06-11T08:12Z [---] followers, [---] engagements


"@futuristflower Most "normies" don't like/don't understand AI. I talk about AI only to selected fews since most ""normies"" tipycally react with a variable mix of fear/disgust/confusion. When you don't (or can't.) understand something one of the most common reaction is fear/opposition"  
[X Link](https://x.com/dieaud91/status/1804538860484940119)  2024-06-22T15:36Z [---] followers, [---] engagements


"@kimmonismus I agree with your "edit" take. I've got a weird feeling about OpenAI. It started in November and it's only getting worse as months pass. It might just be that they are abandoning a sinking (or just heavily decelerating) ship. Government involvement might be the main issue"  
[X Link](https://x.com/dieaud91/status/1820788050173026447)  2024-08-06T11:44Z [---] followers, [---] engagements


"@anissagardizy8 @aaronpholmes @amir Nice. Faster is better. We can't afford to slow down the process of tapping into the real power of AI"  
[X Link](https://x.com/anyuser/status/1843651064698155236)  2024-10-08T13:54Z [---] followers, [---] engagements


"@flowersslop @AnthropicAI @OpenAI @GoogleDeepMind Release the Titans"  
[X Link](https://x.com/dieaud91/status/1844428196294562093)  2024-10-10T17:22Z [---] followers, [---] engagements


"@_Mira___Mira_ ☹☹"  
[X Link](https://x.com/dieaud91/status/1844429731237855678)  2024-10-10T17:28Z [---] followers, [---] engagements


"What truly gets me thinking is how most people are still going about their lives and plans as if the arrival of impactful and powerful AI isnt just around the corner. It feels like anyone outside our niche either doesnt realize whats coming or is simply choosing to ignore it living as if [--] years from now the trajectory of life will be the same as it could have been imagined [--] years ago. The landscape is shifting faster than most are willing to acknowledge; SOTA models are already valuable allies in the work and day-to-day life of those who are aware. Burying our heads in the sand won't make"  
[X Link](https://x.com/anyuser/status/1850214334351171674)  2024-10-26T16:34Z [---] followers, [---] engagements


"@oliviergodement @tarekayed00 @romainhuet Soon 💀💀💀"  
[X Link](https://x.com/dieaud91/status/1851579471129686440)  2024-10-30T10:58Z [---] followers, [----] engagements


"As a developer/tech enthusiast what scares you the most about AI progress I've been thinking a lot about this. Will share my perspective based on votes. Drop your specific concerns below 👇 😱 Being replaced by AI 🚀 Not moving FAST enough 🌍 AI-Society disconnect 🎭 Losing control of AI 😱 Being replaced by AI 🚀 Not moving FAST enough 🌍 AI-Society disconnect 🎭 Losing control of AI"  
[X Link](https://x.com/dieaud91/status/1852347819425665396)  2024-11-01T13:51Z [---] followers, [----] engagements


"@runwayml Mind-blowing how fast this is evolving. Camera movements in AI content used to be sci-fi just months ago. That shot with the girl. The consistency in freckles and hair while panning is insane. Runway keeps pushing boundaries"  
[X Link](https://x.com/dieaud91/status/1852365648900370524)  2024-11-01T15:02Z [---] followers, [----] engagements


"Expanding on point 2: In my profession and discussing with experts from other fields Ive noticed that todays top AI models work best when guided by domain specialists. For example someone without expertise in fields like mine (nutrition) or medicine might miss subtle nuances that make an AI's response subtly incorrect. Non-experts often struggle to ask precise questions that elicit truly professional answers from AI. Thats why at this stage I see current AI models as ideal "coworkers" for experts rather than standalone solutions. We need human expertise to unlock their full potential and"  
[X Link](https://x.com/dieaud91/status/1852406447449223566)  2024-11-01T17:44Z [---] followers, 32.8K engagements


"@tsarnick Imo the Cold War parallel is striking. With AGI/ASI on the horizon we can't afford a single-player game. Balanced power distribution isn't just preferred it's necessary for safety. A 3-10x investment gap could create dangerous asymmetries"  
[X Link](https://x.com/dieaud91/status/1852452736035643595)  2024-11-01T20:48Z [---] followers, [---] engagements


"Fromthe timeline: 2015-2020: A release every year 2023-2024: Multiple releases within months But the real deal isn't the pace it's the compound effect. Each release builds on previous ones exponentially. [----] will make [----] look slow"  
[X Link](https://x.com/anyuser/status/1852814214517444675)  2024-11-02T20:45Z [---] followers, [---] engagements


"Agree with Ilya on this. This is one of the factors that explains why AGI timeline predictions vary so much between experts. A physical body might help but "consciousness" - imo - is mostly related to information processing and the ability to plan and perform actions not embodiment"  
[X Link](https://x.com/dieaud91/status/1853061703765721297)  2024-11-03T13:08Z [---] followers, [---] engagements


"@tszzl Spill the beans roon we need to know"  
[X Link](https://x.com/anyuser/status/1853951719106818307)  2024-11-06T00:05Z [---] followers, [---] engagements


"@rand_longevity C. UBS (Universal Basic Services) + Agency (it's possible to "work" if you want more of something)"  
[X Link](https://x.com/dieaud91/status/1854676091819991096)  2024-11-08T00:03Z [---] followers, [---] engagements


"As a biologist I disagree. Creatine monohydrate is stable in water for several hours. While it will eventually break down into creatinine (an "inactive form") if left in water for an extended period mixing it with water shortly before consumption does not significantly reduce its effectiveness"  
[X Link](https://x.com/anyuser/status/1855377836250472627)  2024-11-09T22:32Z [---] followers, [----] engagements


"@yacineMTB Prove it"  
[X Link](https://x.com/dieaud91/status/1855378205223387351)  2024-11-09T22:33Z [---] followers, [----] engagements


"@btibor91 @teknium I tend to think that the "limit" is the pure "GPT" architecture. Taking into account what @sama and @polynoamial I bet that - internally - this is not a problem anymore. They know how to trascend the "pure GPT""  
[X Link](https://x.com/dieaud91/status/1855383947368157353)  2024-11-09T22:56Z [---] followers, [----] engagements


"Also we are talking about GPTs (and not AI in general) heading to diminishing returns. However OpenAI has been clear that they are pursuing different approaches ("o series") and they seem to be working. We have seen similar patterns in biology and science in general: when a specific approach hits diminishing returns innovation often comes from unexpected directions. "AI models" does not equal GPT"  
[X Link](https://x.com/dieaud91/status/1855628288405733703)  2024-11-10T15:07Z [---] followers, [---] engagements


"@yacineMTB Weekend is for relaxing no need to work"  
[X Link](https://x.com/dieaud91/status/1855675262152909093)  2024-11-10T18:13Z [---] followers, [----] engagements


"@alexalbert__ Most people don't know about Claude. It's a great model especially for writing and "wit" tasks. However too often it refuses to answer without a real reason (example: medical questions I can find on google)"  
[X Link](https://x.com/dieaud91/status/1856398732654174237)  2024-11-12T18:08Z [---] followers, [---] engagements


"@emollick Finally someone getting SoTA performance on consumer hardware. The real revolution isn't in building bigger models but in making them efficient enough to run locally. Qwen [---] could be the beginning of truly personal AI"  
[X Link](https://x.com/anyuser/status/1856486540446183457)  2024-11-12T23:57Z [---] followers, [---] engagements


"@emollick Office integration + agentic AI will be HUGE. Not just about making slides but AI actively running entire workflows through the tools everyone uses daily. Whoever captures this market will win big - most jobs still live in Excel PowerPoint and Word"  
[X Link](https://x.com/dieaud91/status/1857942059866349786)  2024-11-17T00:21Z [---] followers, [----] engagements


"@DaveShapi At some point not using powerful AI for diagnosis might be considered malpractice - just like not using for example X-rays or blood tests would be today"  
[X Link](https://x.com/anyuser/status/1858211590287335599)  2024-11-17T18:12Z [---] followers, [----] engagements


"@emollick Fascinating results and this is "just" GPT-4o. Achieving 85% accuracy in simulating human behavior is massive. This breakthrough could pave the way for AI to tackle highly complex interactions such as personalized healthcare advice"  
[X Link](https://x.com/anyuser/status/1858665529050869898)  2024-11-19T00:16Z [---] followers, [----] engagements


"@btibor91 This makes sense after their emphasis on "agents" and "tool use". Dev day is near we'll know more soon"  
[X Link](https://x.com/dieaud91/status/1859347836342808799)  2024-11-20T21:27Z [---] followers, [---] engagements


"@tszzl They miss both "investment" and basic resource optimization. Using AI to increase productivity/efficiency creates net positive environmental impact. The energy spent on inference is minimal compared to the optimization gains in countless industries. Classic degrowth myopia"  
[X Link](https://x.com/dieaud91/status/1860084575512502447)  2024-11-22T22:14Z [---] followers, [---] engagements


"@emollick @AlecStapp As someone with a biology background this is mind-blowing but also frustrating. $24B is nothing compared to the potential ROI. For context: we spend more yearly on pet food. The tech is there. The science is there. We're just. waiting 🤔"  
[X Link](https://x.com/dieaud91/status/1860085583319826682)  2024-11-22T22:18Z [---] followers, [---] engagements


"@kimmonismus Gemini [----] is quite good. For them to launch [---] it must bring substantial improvements beyond incremental updates"  
[X Link](https://x.com/anyuser/status/1860256519842853009)  2024-11-23T09:38Z [---] followers, [---] engagements


"What many miss is how this cost reduction creates non-linear effects in capabilities. It's not just "faster processing"; each order of magnitude enables completely new applications we couldn't even imagine before. It's not all about parameters. Jensen Huang says NVIDIA has reduced the cost of computing by [--] million times in the past [--] years and although their GPUs are expensive "imagine a million times higher" https://t.co/OKcFdOlnsg Jensen Huang says NVIDIA has reduced the cost of computing by [--] million times in the past [--] years and although their GPUs are expensive "imagine a million"  
[X Link](https://x.com/dieaud91/status/1860428926046044489)  2024-11-23T21:03Z [---] followers, [---] engagements


"@yacineMTB Indeed 3.2k CLEAN calories are definitely a chore"  
[X Link](https://x.com/dieaud91/status/1860699342606102799)  2024-11-24T14:57Z [---] followers, [----] engagements


"@yacineMTB Also you should remember to treat it like a dog and not like a human. It doesn't mean "bad". Just according to his nature (including having enough space to roam and run)"  
[X Link](https://x.com/anyuser/status/1861396277616898551)  2024-11-26T13:07Z [---] followers, [----] engagements


"@Arktophylaxxx Spill the beans man"  
[X Link](https://x.com/dieaud91/status/1861464809956331995)  2024-11-26T17:39Z [---] followers, [---] engagements


"@apples_jimmy I guess Sora will be delayed even more. Nvm personally I'm more in interested in new models (not video)"  
[X Link](https://x.com/anyuser/status/1861543477130969428)  2024-11-26T22:52Z [---] followers, [---] engagements


"@emollick I also felt something "off" in that tweet from @karpathy. Sometimes o1 and Claude give me answers related to my profession that are not what an average random labeler would give. They are what a GOOD colleague would give me"  
[X Link](https://x.com/dieaud91/status/1862644344915661258)  2024-11-29T23:46Z [---] followers, [----] engagements


"@NorthstarBrain @OpenAI I noticed the same. Recently 4o is using search way too often. I don't always want an answer I could find on my own on a search engine. We should have the option to turn search off"  
[X Link](https://x.com/anyuser/status/1862975672831877573)  2024-11-30T21:43Z [---] followers, [----] engagements


"@tszzl Don't tease us like this man. Spill"  
[X Link](https://x.com/anyuser/status/1863883055045235146)  2024-12-03T09:48Z [---] followers, 19.8K engagements


"@slow_developer What if they surprise us with their first attempt with agents Maybe that "Operator""  
[X Link](https://x.com/dieaud91/status/1863912884058624401)  2024-12-03T11:47Z [---] followers, [----] engagements


"@Arktophylaxxx Can't cope with all this hype😵💫"  
[X Link](https://x.com/anyuser/status/1863991162270294272)  2024-12-03T16:58Z [---] followers, [---] engagements


"@apples_jimmy @markchen90 It started as "Open". As time passes it gets more and more closed and we are left with hype and vague hints. Regular progress updates would be appreciated from humanity"  
[X Link](https://x.com/dieaud91/status/1864191725914276186)  2024-12-04T06:15Z [---] followers, [----] engagements


"@emollick Non-experts will have a hard time to feel the difference between powerful models and "legacy" ones. That's why experts will stay relevant at least for some time: they will be the mediators between powerful models and people who are not expert in that field"  
[X Link](https://x.com/dieaud91/status/1864743203401871516)  2024-12-05T18:46Z [---] followers, [----] engagements


"🚨OpenAI o1-pro: This is massive for biology and sciences in general. We've been granted a model that can reason deeply and answer hard questions that would take a human quite some time. We're about to experience an acceleration in sciences. Can't wait to have researchers using o1-pro. #o1pro #ChatGPT #AI #BioAI OpenAI demos ChatGPT o1 pro mode which can think for minutes at a time about complex science and math problems can reason over both text and images and is faster than o1-preview https://t.co/gfMlP6o5j5 OpenAI demos ChatGPT o1 pro mode which can think for minutes at a time about"  
[X Link](https://x.com/anyuser/status/1864746317186240857)  2024-12-05T18:59Z [---] followers, [---] engagements


"@legit_rumors Can't wait to try GPT-4.5. I hope we'll also get a preview of agents"  
[X Link](https://x.com/anyuser/status/1865794231035420690)  2024-12-08T16:23Z [---] followers, [---] engagements


"@Angaisb_ Personal opinion: today we'll get something not that interesting (or at least interesting just for a few). This way they let people talk about Sora. However. What if some other player enters the scene with a big release and steals the thunder 👀"  
[X Link](https://x.com/anyuser/status/1866451983302865257)  2024-12-10T11:56Z [---] followers, [--] engagements


"@Angaisb_ They created so much hype over their next releases that even if what they deliver is objectively great it feels underwhelming. The tech is impressive but not disruptive; we need to manage expectations better and lower them. Too much hype can overshadow real progress"  
[X Link](https://x.com/dieaud91/status/1866548355100516813)  2024-12-10T18:19Z [---] followers, [---] engagements


"🔥🔎 Unfortunate leak or. intentional anticipation from them That's what we're all waiting for. Agents are what will revolutionize the world. https://t.co/Z6q7BGfhsD https://t.co/Z6q7BGfhsD"  
[X Link](https://x.com/anyuser/status/1866580072058265962)  2024-12-10T20:25Z [---] followers, [---] engagements


"@EMostaque Shipping Flash [---] instead of Pro first was a wise choice: - cheaper to run for them - achieves SOTA among small models - showcases a glimpse of [---] family capabilities - dominates the SLM landscape"  
[X Link](https://x.com/anyuser/status/1866947368752742617)  2024-12-11T20:45Z [---] followers, [----] engagements


"@OpenAI "Her" is slowly becoming a reality"  
[X Link](https://x.com/anyuser/status/1867272938002522184)  2024-12-12T18:18Z [---] followers, [----] engagements


"⚠ One detail many are missing from Ilya's quote: "pre-training AS WE KNOW IT will end" As we know it - key words here. This doesn't mean we can't get anything more from it. Rather they'll likely find new approaches and ways to optimize pre-training further while also working on post-training and other aspects. The devil is in the details. Ilya Sutskever says that data is the fossil fuel of AI and "we have achieved peak data" because "we have but one internet" so the age of pre-training AI on bigger and bigger compute clusters "will unquestionably end" https://t.co/mG9jNdATFD Ilya Sutskever"  
[X Link](https://x.com/anyuser/status/1867866557231038580)  2024-12-14T09:37Z [---] followers, [---] engagements


"@kimmonismus My gut tells me no [---]. At least not before [----]. We'll see"  
[X Link](https://x.com/anyuser/status/1868057377309909306)  2024-12-14T22:16Z [---] followers, [----] engagements


"@emollick The Nokia parallel is spot on. They dominated on hardware but failed to see how smartphones would redefine "mobile phone". Now Apple excels in UX but might miss how AI will redefine "smartphone""  
[X Link](https://x.com/dieaud91/status/1868818016307495080)  2024-12-17T00:38Z [---] followers, [----] engagements


"We built AI by looking at our brains. Now AI is helping us look back at ourselves. The ultimate reverse engineering. Demis Hassabis says AI started off inspired by the architecture of the brain and it is now coming full-circle and will help us understand how our brains work https://t.co/02wwSLMRaj Demis Hassabis says AI started off inspired by the architecture of the brain and it is now coming full-circle and will help us understand how our brains work https://t.co/02wwSLMRaj"  
[X Link](https://x.com/anyuser/status/1869487060136174070)  2024-12-18T20:57Z [---] followers, [---] engagements


"@btibor91 Let's hope this isn't what they mean by "agents"; hopefully just a small piece of something much bigger"  
[X Link](https://x.com/anyuser/status/1869515028166238220)  2024-12-18T22:48Z [---] followers, [----] engagements


"Just tried some hard prompts related to my profession (nutrition) and the model nailed the answer. Remember this is Gemini "Flash". Google cooked this time. Let's see if OpenAI has an answer to this. #googlegemini AI Studio users now can access Gemini [---] with COT capability for FREE 🤯 https://t.co/hFfONTw1S7 AI Studio users now can access Gemini [---] with COT capability for FREE 🤯 https://t.co/hFfONTw1S7"  
[X Link](https://x.com/anyuser/status/1869789329977200939)  2024-12-19T16:58Z [---] followers, [---] engagements


"@kimmonismus Just tried this model with some hard prompts and it's good especially considering it is a "flash-class" so a small LM. Let's see if OpenAI has got something worth the wait"  
[X Link](https://x.com/anyuser/status/1869811714264576490)  2024-12-19T18:27Z [---] followers, [----] engagements


""This thing is smarter than me at a lot of hard problems". That's what @sama expect us to feel while interacting with next gen models before the end of [----]. I stay on my prediction: [----] will be the year of the masses awakening to the AI era. Sam Altman says the term AGI gets less useful as we approach it but the o1 model seems close and by the end of [----] we will AI systems that "can do truly astonishing cognitive tasks" where it will seem smarter than humans at a lot of hard problems https://t.co/DIHx3ktrAN Sam Altman says the term AGI gets less useful as we approach it but the o1 model"  
[X Link](https://x.com/anyuser/status/1870042336677290460)  2024-12-20T09:43Z [---] followers, [----] engagements


"@slow_developer I think we'll enter in a period in which AIs become superhuman (asi) in specific fields while below human performance in others. Basically the debate will heat up because of this "jagged frontier""  
[X Link](https://x.com/anyuser/status/1870462407279304747)  2024-12-21T13:32Z [---] followers, [---] engagements


"@Neuralithic @TheGregYang Perhaps it turned out to be an improvement that didn't justify its cost. So.(Speculation) Rebranded Grok [---] and new Training run launched"  
[X Link](https://x.com/anyuser/status/1871584633126900190)  2024-12-24T15:52Z [---] followers, [----] engagements


"I have to admit that I found myself changing perspective about some issues while talking with o1 or Sonnet [---]. The way they present their perspective paired with their style of conversation and their breath of knowledge can sometimes make you reconsider your view (more often than with other humans)"  
[X Link](https://x.com/dieaud91/status/1871607045629087834)  2024-12-24T17:21Z [---] followers, [---] engagements


"DeepSeek-V3 scores 89.1% on MMLU Redux outperforming GPT-4o and Claude [---] Sonnet. Pricing (without discount): - Input: $0.27/M tokens ($0.07 with cache hits) - Output: $1.10/M tokens. As @AILeaksAndNews noted in a comment these powerful cheap open source models are what will reshape the market in the near future 🚀 💰 API Pricing Update 🎉 Until Feb 8: same as V2 🤯 From Feb [--] onwards: Input: $0.27/million tokens ($0.07/million tokens with cache hits) Output: $1.10/million tokens 🔥 Still the best value in the market 🐋 3/n https://t.co/OjZaB81Yrh 💰 API Pricing Update 🎉 Until Feb 8: same"  
[X Link](https://x.com/anyuser/status/1872315726511362401)  2024-12-26T16:17Z [---] followers, [---] engagements


"Which milestone is most likely hit in [----] 🔮 Reliable agentic AI 🤖 Image Generation "Solved" AGI goalposts moved 🤡 OpenSource passes ARC-AGI Reliable agentic AI 🤖 Image Generation "Solved" AGI goalposts moved 🤡 OpenSource passes ARC-AGI"  
[X Link](https://x.com/anyuser/status/1872786712482779559)  2024-12-27T23:28Z [---] followers, [----] engagements


"@kimmonismus Missing out on nuclear in the AI era is like shooting ourslves in the foot. I hope Europe wakes up"  
[X Link](https://x.com/anyuser/status/1873382878870212812)  2024-12-29T14:57Z [---] followers, [---] engagements


"@BenjaminDEKR The business case might lie in building ecosystems rather than direct profitability. "AI" might be thought of as infrastructure: AWS didn't turn a profit immediately either. Monetization may come from partnerships data pipelines or API integrations"  
[X Link](https://x.com/dieaud91/status/1874137389364634082)  2024-12-31T16:55Z [---] followers, [---] engagements


"@mark_k I hope we get Gemini [---] Pro Thinking and at least a preview of "Operator""  
[X Link](https://x.com/anyuser/status/1874453777211957595)  2025-01-01T13:53Z [---] followers, [---] engagements


"@AILeaksAndNews If they all get released in January-February (accelerating progress even further) [----] might be the year when everything truly changed"  
[X Link](https://x.com/anyuser/status/1874488966608752993)  2025-01-01T16:12Z [---] followers, [----] engagements


"@emollick The irony is that while human grad students numbers decline we're about to see millions of digital minds operating at PhD level research capacity"  
[X Link](https://x.com/dieaud91/status/1875672681095553371)  2025-01-04T22:36Z [---] followers, [---] engagements


"@emollick The challenge is that by the time a rigorous social science study gets published the LLMs discussed in it are already 1-2 generations behind. Academia needs to adapt its pace to exponential progress (maybe using AI)"  
[X Link](https://x.com/dieaud91/status/1876004326851870866)  2025-01-05T20:34Z [---] followers, [---] engagements


"@sama "We believe that in [----] we may see the first AI agents join the workforce and materially change the output of companies." Let's hope it's not just companies but also regular plus/pro users"  
[X Link](https://x.com/anyuser/status/1876083063677243650)  2025-01-06T01:47Z [---] followers, [----] engagements


"@yacineMTB As a European I'm more and more concerned about the blind regulator approach that the EU is pursuing. They need to change route FAST or else we'll all be in serious troubles as powerful AIs emerge and the economic landscape changes"  
[X Link](https://x.com/dieaud91/status/1876267123632451932)  2025-01-06T13:58Z [---] followers, [----] engagements


"@DaveShapi As @sama recently suggested. How near are we to the event horizon Are we already past it (internally)"  
[X Link](https://x.com/anyuser/status/1877487986037723493)  2025-01-09T22:49Z [---] followers, [---] engagements


"@slow_developer This is a big deal. If deployed at scale and when "affordable" having a mid-level engineer "in your pocket" means anyone could build their apps/tools. It will be massive for early adopters"  
[X Link](https://x.com/anyuser/status/1877799896671031788)  2025-01-10T19:29Z [---] followers, [----] engagements


"@tsarnick I'm not sure how I feel about models being ". trained to believe they are not conscious". Is it. Right"  
[X Link](https://x.com/anyuser/status/1878190942743667024)  2025-01-11T21:23Z [---] followers, [----] engagements


"@Dr_Singularity Spill the beans bro"  
[X Link](https://x.com/anyuser/status/1878386368667070706)  2025-01-12T10:19Z [---] followers, [----] engagements


"As I said recently society MUST prepare for the AI age. If we keep living the old way ignoring what's about to come we're heading for bad times. "Long-term" is now a 5-year time frame. [--] years into the future is like [--] years now. Sam Altman says he now thinks a fast AI takeoff is more likely than he did a couple of years ago happening within a small number of years rather than a decade https://t.co/I2Cw7gZ9F6 Sam Altman says he now thinks a fast AI takeoff is more likely than he did a couple of years ago happening within a small number of years rather than a decade https://t.co/I2Cw7gZ9F6"  
[X Link](https://x.com/anyuser/status/1879102305322328188)  2025-01-14T09:44Z [---] followers, [----] engagements


"@emollick The fact that [--] weeks of AI tutoring = [--] years of typical learning is staggering. It shows how AI can be a true "human augmenter" and we're talking about GPT-4"  
[X Link](https://x.com/anyuser/status/1879634597295849839)  2025-01-15T20:59Z [---] followers, [----] engagements


"@polynoamial Could you tell us a little bit more to help separate concrete progress signals from the vague hype"  
[X Link](https://x.com/anyuser/status/1880335280730997141)  2025-01-17T19:24Z [---] followers, [----] engagements


"@apples_jimmy Also if it's not better than Sonnet [---] new on all metrics it would be disappointing"  
[X Link](https://x.com/anyuser/status/1880889731892084782)  2025-01-19T08:07Z [---] followers, [---] engagements


"@polynoamial I'm more and more reading "AGI" associated with "surpassing domain experts". Has the definition changed so much Wasn't just a year ago an AI competent as a median human In my opinion an AI as capable as domain experts in all human fields is by definition superhuman"  
[X Link](https://x.com/anyuser/status/1881072973521842261)  2025-01-19T20:15Z [---] followers, [----] engagements


"@Dr_Singularity Even if we don't get "super agents" but just agents reliable/smart enough to replace - let's say - 15% of jobs it'll be disruptive"  
[X Link](https://x.com/anyuser/status/1881100056302064024)  2025-01-19T22:02Z [---] followers, [---] engagements


"@kimmonismus Based on the current trajectory and assuming we don't hit a major roadblock I expect something like o5/o6 to surpass human intelligence. This shouldn't happen any later than 2026"  
[X Link](https://x.com/anyuser/status/1881692455051165865)  2025-01-21T13:16Z [---] followers, [----] engagements


"@emollick Moreover the general public has no idea of what their "AGI" definition (Powerful AI) implies. (Summary by Claude Sonnet)"  
[X Link](https://x.com/dieaud91/status/1881782976390046204)  2025-01-21T19:16Z [---] followers, [---] engagements


"@yacineMTB And what about what we are not discovering soon enough"  
[X Link](https://x.com/anyuser/status/1881814828416991710)  2025-01-21T21:23Z [---] followers, [---] engagements


"@OfficialLoganK Google is cooking. A Flash class model (small 15-30B) achieving great performance in Math and GPQA Diamond. Can't wait to try Gemini Pro"  
[X Link](https://x.com/dieaud91/status/1881845220767322144)  2025-01-21T23:23Z [---] followers, [----] engagements


"@TheMongrel_Cat @AlvaMarcielle As an Italian: well done. Refusing nuclear power in the age of AI would mean becoming irrelevant"  
[X Link](https://x.com/dieaud91/status/1882407023402422525)  2025-01-23T12:36Z [---] followers, [---] engagements


"@rolyataylor2 @flowersslop"  
[X Link](https://x.com/dieaud91/status/1882438063688978826)  2025-01-23T14:39Z [---] followers, [---] engagements


"@tsarnick The timelines are shortening for all AI labs"  
[X Link](https://x.com/anyuser/status/1882525658993295412)  2025-01-23T20:27Z [---] followers, [---] engagements


"@BrentM_SpaceX Thats not th3 upgrad3 w3 ar3 waiting for"  
[X Link](https://x.com/dieaud91/status/1883304084679315824)  2025-01-26T00:00Z [---] followers, [---] engagements


"That's one of the reasons why we need more transparent communication and broader public awareness about what's coming. Here on X we're almost in a "bubble". Most news coverage I see is either outdated - based on technology from a year ago - or consists of doomsday perspectives rooted purely in fear. The top AI Labs should be far more transparent with the public about their internal capabilities. It's concerning that only governments have access to this information and can "hold" it until politically convenient. Society needs to prepare for what's coming - it's truly unprecedented. The public"  
[X Link](https://x.com/dieaud91/status/1883560277200801888)  2025-01-26T16:58Z [---] followers, 10.5K engagements


"Finally Europe makes a move into agentic AI. Do we stand a chance We HAVE to - the alternative is falling further behind. Were excited to announce that Proxy Europes answer to @OpenAI's Operator will be launching globally today with basic access for free and a $20p/m unlimited session plan. https://t.co/RkqDoBTwpx Were excited to announce that Proxy Europes answer to @OpenAI's Operator will be launching globally today with basic access for free and a $20p/m unlimited session plan. https://t.co/RkqDoBTwpx"  
[X Link](https://x.com/anyuser/status/1883895261598298485)  2025-01-27T15:10Z [---] followers, [---] engagements


"@kimmonismus Previous safety researchers' exaggerations (remember the GPT-2 case) leave us with a doubt: while likely in good faith is this just another false alarm. or should we be concerned"  
[X Link](https://x.com/dieaud91/status/1883972638386909592)  2025-01-27T20:17Z [---] followers, 13.4K engagements


"@tsarnick That's it. NVDA's fall actually creates a good swing to buy more stocks"  
[X Link](https://x.com/dieaud91/status/1883973512878321994)  2025-01-27T20:21Z [---] followers, [----] engagements

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