Joined January 2023
raffaelo retweeted
GPT-Rosalind surfaced a new lead for drug discovery: could restoring cholesterol transport protect neurons from degeneration? DCTN1 is a gene linked to Perry syndrome, a rare neurodegenerative disorder. In mouse-brain data, disrupting it increased expression of genes involved in making cholesterol. GPT-Rosalind suggested a possible explanation: cholesterol may not be reaching where neurons need it. That gives scientists a specific mechanism to test and, if confirmed, a direction for developing new treatments for neurodegenerative disorders. With @perturbai_tx and @scrippsresearch, a team of thousands of GPT-Rosalind agents worked in parallel, connecting gene-expression results from millions of mouse brain cells with scientific literature in just 32 minutes. We’re now expanding this collaboration to identify new drug targets. openai.com/index/scaling-ai-…
377
668
280
6,433
1,236,864
OpenAI is under INSANE pressure to deliver. My bet? They'll drop a new GPT Sol or Astra in the next few weeks and hype it like they've finally cracked AGI. The demos will be absolutely bonker, we'll all lose our shit. Then we'll get our hands on it and find out it's as mediocre or useless as it comes...or not.
45
1
68
3,912
I have an AI agent that logs every negative comment someone makes about me online. It then creates a profile of the person, maps all of their social connections, and finds all their emails. Tomorrow morning, my AI agent will sent out an email blast to everyone's extended social network, detailing the horrible things they've said about me.
1,999
499
434
25,619
3,125,632
Elon Musk stutters. Not occasionally. Consistently. In almost every interview and public appearance, there are moments where he pauses mid sentence for 3 to 5 seconds, searching for a word that his brain has processed but his mouth cannot produce. The pauses are long enough that interviewers visibly shift in discomfort. The audience holds its breath. The man who runs 6 companies simultaneously cannot always produce the next word in a sentence. The speaking difficulty is not a performance. It is not a thinking pause (though he has those too). It is a genuine processing bottleneck between his thoughts and his speech. His brain operates faster than his verbal output system can manage. The thoughts are there. The words lag. The lag produces the stutter, the filler sounds, and the long silences that define his public speaking. As a child in South Africa, the speaking difficulty was more pronounced. He was slow to speak. He was quiet in class. Teachers assumed he was intellectually behind. The boy who was reading the Encyclopaedia Britannica at 9 was perceived as struggling because his verbal output didn't match his cognitive input. The gap between what he knew and what he could say was the first version of a pattern that would follow him for life. The irony is devastating: the man whose companies depend on public trust, investor confidence, and media communication is physiologically disadvantaged at the one activity that produces all three. Every keynote, every earnings call, every interview requires him to perform the skill his nervous system handles worst. He does it anyway. Poorly. Consistently. Without avoiding it. He has never used a teleprompter. He has never used scripted remarks. He speaks live, stutters live, pauses live, and recovers live. The refusal to script is not bravery. It is authenticity pushed to a point that makes his communications team visibly nervous. The man would rather stutter through the truth than fluently deliver a script. The most powerful communicator of the 21st century, the man whose tweets move markets and whose vision has redirected entire industries, cannot always finish a sentence without a 5 second pause. The words don't come easy. They never have. He says them anyway. And the world that watches him stumble through a keynote then watches him land a rocket and concludes that the stumble is what makes the landing believable. The man is not polished. He is real. The stutter is the proof.
309
307
42
2,866
161,374
raffaelo retweeted
Day 16/ You can now cache for 12h in the API and for logged in users across Codex & Work we keep cache forever and you don’t have any penalty for coming back to an old conversation or resuming after a few hours and are now guaranteed to hit the cache.
You can now choose 12-hour prompt caching in the API. Persistent agents, like dots, reuse context throughout the day. You can keep shared instructions, tools, and context cached longer and pay the lower cached-input rate, even when requests are hours apart.
2,575
680
1,073
15,287
2,789,692
I hit four years at OpenAI this weekend. Some thoughts - OpenAI changed a lot this year - Years 1 and 4 were the hardest, but for different reasons - Not many people left from when I joined but still so many amazing ppl Not exactly sure how i'm feeling tbh but it's been cool
18
7
526
45,213
raffaelo retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 github.com/deepseek-ai/DeepI… 1/4
1,544
4,747
2,287
42,059
6,367,148
raffaelo retweeted
Fable 5.5 is apparently dropping tomorrow now. yesterday it was today. last week it was also today. new reports suggest Anthropic could be getting ready, and there are signs of partner testing. still no official announcement though. meanwhile, GPT-6.1 Terra is showing up in coding evaluation references, and OpenAI hasn’t said a word about it either. we’re getting more model leaks than actual models at this point
15
2
1
69
12,646
raffaelo retweeted
An Anthropic researcher just told a room full of PhD students that their career plan no longer exists His 3 pieces of advice for students: 1. Use AI every day. If you're on the exponential, you'll never fall behind it. 2. Choose projects that get better when compute scales. The rest will be solved before you finish them. 3. Stop planning your career as if it were stable. The safe option no longer buys safety; it no longer exists. And a phrase that sums it all up: "Writing is thinking. If you outsource your writing, you outsource your thinking." The talk is titled "The End of the Anthropocene"
37
190
29
1,579
224,475
raffaelo retweeted
There were reports of usage draining faster than anticipated for some users over the weekend. We've identified approximately 40,000 accounts that had an inconsistency on which usage limits where the actual usage that was applied did not correspond to their actual selected plan. We are reaching out individually to all these users and have corrected the issue. If this affected you, we'll add credits to your account. Sorry we got this wrong.
2,823
356
458
11,853
1,746,428
Introducing SKY-001, the first Engineering World Model. On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested: -> 51-125x faster - > 150-210x cheaper -> Higher IOU than all frontier models SKY-001 lets you dream a physical object before you build it. 18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch). The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.
823
2,397
644
16,366
4,154,560
raffaelo retweeted
Day 15/ Nevermind, we're still cooking... and reset. Enjoy.
Day 15/ We're cooked
2,844
462
885
11,652
1,657,493
raffaelo retweeted
When Julius Caesar was 25, he was kidnapped by pirates. He acted like they worked for him. They demanded a ransom of 20 talents. Caesar laughed at them for not knowing who they had captured, and told them to ask for 50. He spent 38 days as their prisoner. When he wanted to sleep, he ordered them to stop talking. He wrote poems and speeches and read them out loud. When the pirates didn’t admire them, he called them illiterate barbarians. He often told them, laughing, that he would come back and execute them all. They thought he was joking. As soon as the ransom was paid, he raised a fleet, hunted them down and had every one of them crucified, exactly as he had promised.
205
1,213
77
23,265
1,174,802
NEVER try to make people like you. You could LITERALLY build a flying car and donate $10 billion to children, and people will STILL find ways to shit on you. Just be you. DON'T WORRY ABOUT WHAT OTHER PEOPLE THINK! JUST DO COOL SHIT!!!!!
183
356
33
4,875
78,582
Google has another banger 🤯 It’s called PaperBanana. It turns your methodology text into academic illustrations, helping you explain your research visually. 100% free.
16
308
12
2,312
102,874
Having Fable 5.5 be a major jump on some of these would be fantastic, but the real spear in the side of OpenAI would be removing the Fable Stress. So basically having Fable work just like Opus does now, but maybe use your limits a bit faster. But nothing dramatic. And without its own dedicated limit. Double spears and a Thanos snap would be if they did this while raising limits somehow to reduce even more anxiety. But that's just wishful thinking. Either way, can't wait to see it.
Fable 5.5 probably definitely tomorrow. Rumors: - Better at reasoning and intuition - Infers intent rather than taking instructions literally - More ambitious. Ask it to fix a bug and it fixes the deeper root cause - Does creative, “out-of-the-box” thinking - Can invent algos to predict stocks - Bestows complete and eternal enlightenment on first response AND I’d bet it has major efficiency gains like Opus 5.5. Icing on the cake would be usage for all subs. Prepare yourself in mind, body and spirit.
11
6
80
34,813
Fable 5.5 probably definitely tomorrow. Rumors: - Better at reasoning and intuition - Infers intent rather than taking instructions literally - More ambitious. Ask it to fix a bug and it fixes the deeper root cause - Does creative, “out-of-the-box” thinking - Can invent algos to predict stocks - Bestows complete and eternal enlightenment on first response AND I’d bet it has major efficiency gains like Opus 5.5. Icing on the cake would be usage for all subs. Prepare yourself in mind, body and spirit.
159
92
22
3,480
297,715
Diversification is overrated. You shouldn’t avoid risk concentration. Everything important in life demands concentrated risk. All of the most important areas of my life—my marriage, my close friendships, my business partners, my professional skills, and my spiritual life—required me to make choices. When you realize that you have to focus your attention and time it makes you take your choices more seriously. The key to success isn’t hedging your bets. It is picking the right places to go all in.
6
7
1
46
2,677
The people panicking about AI are the same ones who said: • The internet was a fad • Social media was for kids • Remote work would never last They're not wrong because they're dumb. They're wrong because they're comfortable. Comfort is the enemy of the next era.
55
13
5
126
5,165
raffaelo retweeted
Spent the weekend trying out LLM psychosis I had Astra, Opus 5.5, and Argon do science autonomously and try to write research papers Super fun and addictive, feels like you're getting so much done. But then I read the papers and they were almost completely nonsense
56
35
5
1,496
53,634