Alejandro Catalá Espí

107 posts

Alejandro Catalá Espí banner
Alejandro Catalá Espí

Alejandro Catalá Espí

@alejandrocatala

Doctor, pero no de los que cura. Profesor innovador. Padre por partida doble.

Valladolid, Spain Katılım Ağustos 2009
272 Takip Edilen115 Takipçiler
Jon Hernandez
Jon Hernandez@JonhernandezIA·
HOLY COW!! You need to try this... The speed is insane and it totally blows your conception of an Ai for so many reasons...
Jon Hernandez tweet media
English
18
3
52
9.8K
Charrita 🌾
Charrita 🌾@3charrita·
@cyltv Buenas noches, ¿Está en algún sitio el programa "espacio abierto" dónde se entrevistó a @alejandrocatala sobre el eclipse solar? Fue muy interesante. Felicidades por el trabajo y gracias anticipadas por la respuesta. Un saludo,
Español
1
0
0
33
François Chollet
François Chollet@fchollet·
If you want your young kids to practice reading, give them videogames where reading dialogues is necessary to progress.
English
70
37
656
44.5K
Jegap
Jegap@jungsungchacn·
This should be mandatory for all public toilets.
English
126
55
1.6K
1.7M
Shanaka Anslem Perera ⚡
Shanaka Anslem Perera ⚡@shanaka86·
THE PHYSICS OF MONEY JUST CHANGED FOREVER For 10,000 years, humanity faced one impossible problem: Energy cannot be stored. Energy cannot be transported. Energy dies the moment it is born unused. Until now. Jensen Huang, CEO of the $4.5 trillion company that powers artificial intelligence, just stated what changes everything: “Bitcoin is taking excess energy and storing it as a new form called currency. You take that currency wherever you’d like. You took energy from one place and transported it everywhere.” Read that again. The numbers are staggering. In 2024 alone, Texas was forced to throw away 8 terawatt hours of wind and solar energy. Brazil discarded 28 terawatt hours in eight months. Globally, over $20 billion in clean energy vanishes annually because grids cannot absorb it. Bitcoin mining now consumes 211 terawatt hours per year. 52.4% from renewable and nuclear sources. Miners co-locate at stranded energy sites, converting worthless surplus into globally transferable value. This is not digital gold. This is something physics has never permitted before. For the first time in human history, energy has become portable. Joules transformed into hashes. Hashes settled into satoshis. Satoshis moved across borders at the speed of light. The sun sets in Arizona. That captured sunlight arrives in Tokyo as money. Wind dies in West Texas. That momentum resurfaces in London as settlement. Critics said Bitcoin wastes electricity. The man whose chips process most of the world’s AI workloads just told you the opposite: Bitcoin is the battery that never existed. The transmission line that needs no wire. The export of energy without the tanker. The implications are civilizational. Stranded renewables become profitable. Remote regions become energy exporters. The economics of electricity generation inverts permanently. Energy is no longer trapped. Energy is now free.
English
513
364
1.4K
180.1K
Esha
Esha@EshaAA33·
A woman's strange behavior during a press conference in the Oval Office went viral... Who is she? 🤔
English
1.4K
232
1.1K
1.4M
LadyValor
LadyValor@lady_valor_07·
Apparently if you’re right-brained you see a rabbit, if you’re left-brained you see a turtle. What do you see
LadyValor tweet media
English
36.8K
1.4K
26K
7.9M
Alejandro Catalá Espí
Alejandro Catalá Espí@alejandrocatala·
Si lanzas una red neuronal vacía al mundo, no es un niño; es una ameba intentando aprender cálculo.
Español
0
0
0
14
Andrew Ng
Andrew Ng@AndrewYNg·
Releasing a new "Agentic Reviewer" for research papers. I started coding this as a weekend project, and @jyx_su made it much better. I was inspired by a student who had a paper rejected 6 times over 3 years. Their feedback loop -- waiting ~6 months for feedback each time -- was painfully slow. We wanted to see if an agentic workflow can help researchers iterate faster. When we trained the system on ICLR 2025 reviews and measured Spearman correlation (higher is better) on the test set: - Correlation between two human reviewers: 0.41 - Correlation between AI and a human reviewer: 0.42 This suggests agentic reviewing is approaching human-level performance. The agent grounds its feedback by searching arXiv, so it works best in fields like AI where research is freely published there. It’s an experimental tool, but I hope it helps you with your research. Check it out here: paperreview.ai
Andrew Ng tweet media
English
248
1.1K
6.3K
1.1M
Alejandro Catalá Espí
Alejandro Catalá Espí@alejandrocatala·
@NotebookLM Do you have an official prompting guide for generating slide decks in different styles, colour palettes, target audience, etc?
English
0
0
0
12
NotebookLM
NotebookLM@NotebookLM·
Next up… Slide Decks! Turn your sources into a detailed deck for reading OR a set of presentation-ready slides. They are fully customizable, so you can tailor them to any audience, level, and style. Officially rolling out to Pro users now (free users in the coming weeks)!
English
218
827
7K
1.6M
SolidKroket
SolidKroket@SolidKroket·
Necesito una película donde el villano realmente gane.
Español
3.7K
382
9.6K
1.2M
Dios
Dios@diostuitero·
Venga, confesad. En mi caso, Ulises, de James Joyce.
Dios tweet media
Español
1.6K
47
1.1K
155.4K
Hannibal999
Hannibal999@Hannibal9972485·
Basically a taxpayer-funded walled garden, where the returns go to the military-industrial-tech complex, not the people who paid. 🚨🚨 a taxpayer-funded monopoly engine Even though the government is building and securing the Genesis stack, private partners are going to profit from it massively according to documents •They’ll get privileged access to the best datasets and compute. •They’ll co-own any commercially valuable IP produced through the platform. •They’ll get early access to discoveries with military, pharmaceutical, or industrial value. the documents say They’re bringing in private contractors, cloud providers, compute vendors, and AI model builders who will: •Gain access to taxpayer-funded data •Train models on it •Commercialize the results •Gatekeep who else can use them So publicly funded research gets converted into private intellectual property, and then resold back to the government or licensed to select companies. And that’s assuming it stays peaceful. But what if the same models used for drug discovery are also used for: •Designing bioweapons? •Optimizing propaganda? •Building “digital twins” of adversary populations for psychological targeting? It’s already happening in adjacent DARPA projects. Genesis simply connects the plumbing across agencies and contractors. ————- 🚨also It’s quietly laying the rails for state-corporate monopolies over biology, energy, and materials Look carefully: Genesis isn’t just about research. It’s also about manufacturing AI-directed production, autonomous experimentation, simulation-to-fab loops. That means: •The same system that generates a scientific breakthrough will also own the infrastructure to produce it, license it, and deploy it. •And the IP rules baked into the EO already point to government-corporate ownership of those breakthroughs. You will end up with Genetically engineered materials or drugs owned by state-contracted corporations. •New energy designs (fusion, storage, quantum) being patented and licensed from a government-run AI platform. •Whole fields of biology, chemistry, or physics locked up behind legal firewalls, enforced by a model that only they can run. So even if the science “helps humanity,” it will do so on someone else’s terms. You won’t own the cure. You’ll rent it if you’re allowed to. —- 🚨🚨What’s worse Genesis isn’t just for science. It’s a prototype for AI-managed institutions. Watch closely it sets a precedent. Once you show you can run science this way, the logic spreads: •Education: AI determines curriculum, tests, and “truth.” •Healthcare: Diagnosis, treatment, and drug development all shaped by AI systems trained on locked datasets. •Law: Predictive justice systems trained on biased or selective data. What starts as a science initiative becomes the blueprint for a new kind of top-down, simulation-based governance, where everything important is run through centralized, predictive models. And once it’s normalized, opting out won’t be an option. ——- Even worse 🚨 It gives the government long-term leverage over every private innovator Once you plug in whether you’re a startup, lab, university, or corporation you are: Feeding the system your data •Training its models •Building tools inside its infrastructure In return, you get limited access, maybe some compute credits, maybe “preferred partner” status. But the moment your company starts to build something valuable outside the system, guess what happens? •Your access gets pulled. •Your compute dries up. •Your reputation inside the ecosystem tanks. ——— Lastly even WORSE 🚨 It buries danger zones like synthetic biology and AI-directed materials under vague language This system can be used to: Design new pathogens faster than any lab in the world Simulate how they’d spread Optimize their effects They can make Bio-manufactured organisms trained on your knowledge and you won’t know its being unleashed
English
5
25
52
4.8K
Computer
Computer@AskPerplexity·
🚨The White House just launched the Genesis Mission — a Manhattan Project for AI The Department of Energy will build a national AI platform on top of U.S. supercomputers and federal science data, train scientific foundation models, and run AI agents + robotic labs to automate experiments in biotech, critical materials, nuclear fission/fusion, space, quantum, and semiconductors. Let’s unpack what this order actually builds, and how it could rewire the AI, energy, and science landscape over the next decade:
Computer tweet media
English
836
1.6K
9.7K
9.2M
AshutoshShrivastava
AshutoshShrivastava@ai_for_success·
Official: Google has dropped their agentic IDE AntiGravity. I was one of the early testers and this thing is crazy good. Thanks to the DeepMind team for early access. It can create images, control the browser, and do things no IDE can do right now. 5 use cases below 👇 1. Last Sunday I hit an issue with my Supabase setup. It tested my app, found the issue, accessed my Supabase, made the required changes, and fixed everything. Incredible. Some video sources are from Google.
English
40
83
869
164.9K