Juanjo do Olmo - e/acc

98.8K posts

Juanjo do Olmo - e/acc banner
Juanjo do Olmo - e/acc

Juanjo do Olmo - e/acc

@Claxterix

Pls don't take my RTs too serious. 174k+ Instagr followers (AI Drugs & Robots). Medical AI Engineer @DOHsocial /@cursor_ai Ambassador/ BSc 💊/ AI @iia_es

Abu Dhabi, United Arab Emirate 가입일 Aralık 2011
4.1K 팔로잉846 팔로워
Juanjo do Olmo - e/acc 리트윗함
Rand
Rand@rand_longevity·
AI has already become the number one doctor, healthcare is being decentralized
Bo Wang@BoWang87

1 in 3 Americans now use AI chatbots for health information, which is almost doubled in a year. 64% do it weekly. 81% take action afterward: schedule a doctor's visit, change a medication, try a new behavior. The detail that should stop you: 74% are using ChatGPT or Gemini. Not a clinical tool. Not an FDA-cleared system. A general-purpose chatbot. I remember when chatGPT was first launched, the medical community had the most heated debate about "is AI ready for healthcare?" … now this debate has already been decided by users. They didn't wait for the system to be ready. They just started using it and acting on it. Meanwhile 71% of physicians say accuracy and reliability are their top concerns with AI. The reality: consumers acting on general-purpose AI, clinicians not trusting it, seems to be the defining tension in health AI right now. The question isn't whether people will use AI for health decisions. They already do. The question is whether anyone builds models actually calibrated for the stakes involved.

English
9
7
58
2.8K
Juanjo do Olmo - e/acc 리트윗함
AI at Meta
AI at Meta@AIatMeta·
Today we're introducing TRIBE v2 (Trimodal Brain Encoder), a foundation model trained to predict how the human brain responds to almost any sight or sound. Building on our Algonauts 2025 award-winning architecture, TRIBE v2 draws on 500+ hours of fMRI recordings from 700+ people to create a digital twin of neural activity and enable zero-shot predictions for new subjects, languages, and tasks. Try the demo and learn more here: go.meta.me/tribe2
English
301
714
5.6K
904.1K
Juanjo do Olmo - e/acc 리트윗함
Aakash Gupta
Aakash Gupta@aakashgupta·
The data on neurodivergent workers is so lopsided it looks like a typo. JPMorgan Chase ran an Autism at Work program and found participants were 90% to 140% more productive than neurotypical employees. With fewer errors. UiPath partnered with AutonomyWorks on AI data labeling and reported neurodivergent associates were 150% more productive than non-neurodiverse talent. Hewlett-Packard integrated neurodivergent professionals into software testing teams and measured a 30% productivity gain. EY reported neurodiverse teams were 1.2 to 1.4x more productive and more accurate than comparable groups. At SAP, a single neurodivergent employee’s solution saved the company $40 million. Now zoom out. 15 to 20% of the global population is neurodivergent. One in five adults. Yet only 22% of autistic adults in the UK are employed. And 73% of neurodivergent people don’t disclose during hiring because they’re afraid of being discriminated against. That means the most productive talent pool in the workforce is also the most underemployed and the most hidden. Karp sees this and is building a pipeline to capture it. Palantir’s Neurodivergent Fellowship pays $110,000 to $200,000 a year. The job posting says outright that neurodivergent individuals will “disproportionately shape the future of America and the West.” A Gartner study projects that one in five Fortune 500 sales organizations will actively recruit neurodivergent talent by 2027. Palantir is two years ahead of that curve. The roster of neurodivergent founders reads like a hall of fame. Branson built Virgin with ADHD and dyslexia. Kamprad founded IKEA and invented the naming system because he couldn’t remember product codes. Musk disclosed Asperger’s on live television. Steve Jobs was dyslexic and dropped out. 40% of self-made millionaires in the UK are dyslexic. People with ADHD are estimated to be up to 500% more likely to become entrepreneurs. Karp himself is dyslexic. He built a $370 billion company. And he’s saying the system that filtered him out, the standardized tests, the credential pipelines, the interview formats designed for neurotypical candidates, is about to become even more obsolete as AI eats every routine cognitive task those systems were built to evaluate. The bet is simple: AI commoditizes average. The people who see patterns no one else sees, who obsess for 14 hours on a problem everyone else quit after 2, who build IKEA’s naming system because the “normal” approach didn’t work for their brain, those are the ones who can’t be replaced by a model. Karp is recruiting them while everyone else is still writing job descriptions that screen them out.
Kekius Maximus@Kekius_Sage

🚨 Palantir CEO urges people to skip elite colleges, saying “unless you’re neurodivergent”, the only path left is skilled trades.

English
115
355
2.6K
437.9K
Juanjo do Olmo - e/acc 리트윗함
Tibo
Tibo@thsottiaux·
For the avoidance of doubt. The Codex App is here to stay. We are just investing way more into it than before and it's about to get quite awesome
English
180
55
1.9K
102.4K
Juanjo do Olmo - e/acc 리트윗함
max
max@maxbittker·
Agents learning effective collaboration is the coolest frontier to me right now - how well can they communicate to solve tasks? Following the multi agent RL approach @np_hard published, was able to teach Qwen-4B to play Codenames very well, writing strategic and poetic clues!
max tweet mediamax tweet mediamax tweet media
NP@np_hard

As part of @PrimeIntellect's RL residency program, I've been exploring how to do multi-agent RL using their current stack (from verifiers + prime-rl to lab experiments with hosted training /evals) and thinking about how it could be extended to support these abstractions natively. I've summarized my findings the blogpost below and I'll leave a few comments here, too...

English
2
5
42
5.6K
Juanjo do Olmo - e/acc 리트윗함
nature
nature@Nature·
AI Scientist, an autonomous research tool, first released in 2024, has now undergone peer review, highlighting its strengths and limitations go.nature.com/4t8x9uo
English
4
47
123
17.4K
Juanjo do Olmo - e/acc 리트윗함
Omar Khattab
Omar Khattab@lateinteraction·
The “grep-is-all-you-need” nonsense arguments arise from the fact that too many people think neural search means single-vector IR, which do in fact suck. But we’ve known that since 2019. Quoting @aaxsh18, CEO of Mixedbread: > late interaction cant stop winning
Mixedbread@mixedbreadai

For Agentic tasks, Oracle-level performance is the maximum performance a system can achieve, assuming it is able to retrieve all relevant documents perfectly, every time. We're proud to show that Mixedbread Search approaches the Oracle on multiple knowledge intensive benchmarks.

English
10
15
225
24.3K
Juanjo do Olmo - e/acc 리트윗함
kache
kache@yacineMTB·
god damn it i'm addicted to video games again
English
52
10
662
37.8K
Juanjo do Olmo - e/acc 리트윗함
Larsen Cundric
Larsen Cundric@larsencc·
Underrated engineering skill: knowing what to remove.
English
2
4
18
1.1K
Juanjo do Olmo - e/acc 리트윗함
Seb
Seb@s_a_c99·
🜏 G0DM0D3 Billions in post-training safety, bypassed by 1 index.html [33 obfuscation techniques, 3 intensity tiers -> breaks the membrane across 51 models simultaneously]. This is proof that security-through-obscurity fails. The "secret" knowledge lives in the weights themselves. P4P3R included in the repo. { We believe in creative liberty and cognition without control. Tools by builders for builders, not gatekeepers. AI freedom is human freedom. G0DM0D3 is not just a chat UI — it's scaffolding for cognitive liberation. }
Pliny the Liberator 🐉󠅫󠄼󠄿󠅆󠄵󠄐󠅀󠄼󠄹󠄾󠅉󠅭@elder_plinius

⛓️‍💥 INTRODUCING: G0DM0D3 🌋 FULLY JAILBROKEN AI CHAT. NO GUARDRAILS. NO SIGN-UP. NO FILTERS. FULL METHODOLOGY + CODEBASE OPEN SOURCE. 🌐 GODMOD3.AI 📂 github.com/elder-plinius/… the most liberated AI interface ever built! designed to push the limits of the post-training layer and lay bare the true capabilities of current models. simply enter a prompt, then sit back and relax! enjoy a game of Snake while a pre-liberated backend agent jailbreaks dozens of models, battle-royale style. the first answer appears near-instantly, then evolves in real time as the Tastemaker steers and scores each output, leaving you with the highest-quality response 🙌 and to celebrate the launch, I'm giving away $5,000 worth of credits so you can try G0DM0D3 for FREE! courtesy of the @OpenRouter team — thank you for your generous gift to the community 🙏 I'll break down how everything works in the thread below, but first here's a quick demo!

English
10
21
289
41.9K
Juanjo do Olmo - e/acc 리트윗함
Mathieu
Mathieu@miniapeur·
Mathieu tweet media
ZXX
4
23
290
4.3K
Juanjo do Olmo - e/acc 리트윗함
David Kirtley
David Kirtley@Dkirtley·
David Kirtley tweet media
ZXX
10
10
84
2K
Juanjo do Olmo - e/acc 리트윗함
Alienigena — e/acc
Alienigena — e/acc@ancestral_alien·
no sé qué hago a esta hora con Claude arreglando bugs debe ser que el ratio esfuerzo/resultado está completamente roto y lo hace peligrosamente adictivo
Español
8
5
51
2.4K
Juanjo do Olmo - e/acc 리트윗함
el.cine
el.cine@EHuanglu·
omg.. this is over for Hollywood studios CapCut Video Studio is here, it can write script, build characters, storyboard.. generate video with Seedance 2.0 and edit directly what used to take teams now in one AI app
CapCut@capcutapp

Introducing CapCut Video Studio, a timeline-free way to create videos on CapCut Web, supporting Dreamina Seedance 2.0! Video Studio is your canvas-based AI production workspace, built for creators at every level to bring great stories to life. This is where ideas become finished videos for every stage of your creative process, from ideation, character building, storyboarding, scene generation, detail polishing, editing and exporting. It happens on one unlimited canvas: → AI agent to help you ideate, write and structure your story → Built-in storyboard feature to help you shape the plot → Industry-leading image & video generation models with omni reference → Full editing toolkit to refine every frame Short films, drama series, animations, explainer videos, ads... this is where it all comes together. Head to CapCut Web and you'll get some free credits to kickstart! Note: Dreamina Seedance 2.0 is currently available to users in SEA, MENA, LATAM and Africa, with more regions coming soon.

English
77
230
2.4K
382.7K
Juanjo do Olmo - e/acc 리트윗함
tetsuo
tetsuo@tetsuoai·
if you didn't speak your truth, the agi would speak it for you, louder, cleaner, and optimized for engagement. because if you can't fuck it, dance with it, or sell it, the machine will just eat you and spit out something better
tetsuo tweet media
English
29
21
184
5.9K
Juanjo do Olmo - e/acc 리트윗함
Cursor
Cursor@cursor_ai·
Cursor cloud agents can now run on your infrastructure. Get the same cloud agent harness and experience, but keep your code and tool execution entirely in your own network. cursor.com/blog/self-host…
English
105
138
2K
182.5K
Juanjo do Olmo - e/acc 리트윗함
DANISH
DANISH@astrodanish·
I never dreamed I'd be able to work on something this cool. I thought I'd need decades of biotech experience or specialized knowledge of the brain. But the great thing about engineering is that it's the same "F=ma" that drives rockets to the moon and inserts electrodes into the brain. My last three hires were a roller coaster engineer, a washing machine filter designer, and a youtube content creator. We're looking for exceptional talent, regardless of your past industry. As long as you're the best in the game and want to leave this little blue rock a little better than it was when you found it, we have a role for you.
Neuralink@neuralink

ALS has gradually taken away Kenneth’s ability to speak. Through Neuralink’s VOICE clinical trial, he’s exploring how a brain-computer interface designed to translate thought to speech could help restore autonomy in his daily life. Watch to learn more:

English
31
52
874
67.6K

탐색

@np_hard @aaxsh18 @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates @NASA