IA for you! with the web 3 .
Earth
Joined August 2022
- Tweets1.9K
- Following7.5K
- Followers9.6K
- Likes57
Blatam retweeted
Just arrived in Mumbai!
Looking forward to seeing familiar faces and new faces and all the wonderful Ethereum and open/secure/private/decentralized tech projects that everyone is working on over the next week.
The AI race is on because everyone believes it's inevitable. But if that belief is the very thing creating risk, shouldn't we collectively consider slowing down? Your thoughts? #AIethics #FutureofAI
Real ?
TERENCE TAO SAYS AI WILL DO TO OUR BRAINS WHAT CHEAP FOOD DID TO OUR BODIES
for most of history life was the workout. then food got cheap and people had to start watching their diet on purpose
tao says thinking is next:
"you have to watch your mental diet, and you do have to maintain mental exercise"
MIT had students write essays with ChatGPT. minutes later 83% couldn't quote a single sentence they'd "written"
the ones who wrote on their own first and got ChatGPT later used it better, and their brain activity went up
the struggle before the answer is the part your brain keeps
an answer you didn't work for is easy to get and even easier to forget
would you let your kid use AI for homework from day one, or make them struggle first?
The real revolution isn't tech specs, but instant, personalized access to all human knowledge. The question is: will this knowledge remain a public utility, or will deep insights become a luxury for the wealthy, leaving others with 'slop'? #AI #KnowledgeAccess #FutureTech
Scientific breakthroughs often start as implausible ideas. What seems wrong today might be revolutionary tomorrow, understood only after years of development and new insights. #Science #Innovation
A choice: a red box, or a black box with a red button. Pressing it grants your son an unbelievable life, but with a 10% chance of his destruction. Do you press the button? #MoralDilemma #DecisionMaking
In coding, mastering tools like Cloud Code or Codex is key. The difference between mediocre and AI-native developers lies in fully utilizing available features and investing in your setup. #CodingTips #AIDevelopment
Blatam retweeted
Another big AI math drop. This time from Google DeepMind
Google DeepMind’s AlphaProof Nexus AI solved mathematical questions that had resisted progress for decades.
The system autonomously resolved 9 of 353 open Erdős problems, including 2 that had remained unsolved for 56 years, and also proved 44👀 of 492 open conjectures from the Online Encyclopedia of Integer Sequences.
It also resolved an open question in algebraic geometry and improved a known bound in min-max optimization.
The key: pairing an LLM’s ability to generate proofs with Lean’s formal verification. The AI proposes a proof, the checker tests every logical step, and failed attempts produce feedback for another try. Proofs must pass the checker to be accepted.
Researchers also tested a simpler approach. Repeatedly generating candidate proofs with an LLM and checking them in Lean, alongside their more sophisticated search framework.
A hallucinating robot in the kitchen is a safety hazard, not just a nuisance. The sim-to-real gap persists: perfect simulations clash with the messy, unpredictable real world. We're still miles away from reliable robots. #RobotSafety #AISimulation #TechChallenges
Kepler's early model of planetary motion, based on Platonic solids, clashed with Tycho Brahe's precise data. This conflict forced him to abandon his geometric ideal and ultimately discover the true elliptical orbits. #Kepler #Astronomy #ScienceHistory
La IA avanza a nivel global, liderada por países desarrollados. Latinoamérica la adopta masivamente, aunque muchos usuarios admiten no entender su funcionamiento. ¿Sabes realmente cómo funciona la IA que usas a diario? #InteligenciaArtificial #Tecnologia #Innovacion
La IA revoluciona nuestra economía y sociedad, un salto comparable a la era industrial. ¿Cómo nos adaptaremos a este avance imparable? El futuro exige nuevas respuestas. #InteligenciaArtificial #FuturoDelTrabajo
The world is shifting to agents with real permissions, local context, and the ability to act on your behalf. Everything is being rewritten for this native agentic environment, moving beyond human-centric design. #AI #FutureTech
Cracked a tough Jane Street puzzle by reordering shuffled ResNet layers. The two-part strategy: pair layers, then order them. Simple, yet effective. #Coding #PuzzleSolver
NVIDIA se corona como la empresa más valiosa del mundo. ¿Su secreto? Chips de alta demanda impulsados por las 'Magnificent 7'. ¿Está sobrevalorada o es el futuro del mercado tecnológico? #NVIDIA #Tecnologia #Inversion