Carles Lázaro Costa

6.2K posts

Carles Lázaro Costa banner
Carles Lázaro Costa

Carles Lázaro Costa

@Carleslc

👨‍💻 Software Engineer 🤖 🏛️ Filosofía 🔭🧠 Ciencia 📈 Finanzas ⛩️🎎 Anime ⛰️👟 Trail Running 🎲🎭 Rol & Boardgames 💀 Memento mori ☀️ Hakuna matata 🕊️

Barcelona (Laniakea) 🌌 Katılım Aralık 2010
707 Takip Edilen159 Takipçiler
Carles Lázaro Costa retweetledi
Alan
Alan@bitforth·
Yo fui ingeniero en Meta, y siempre seguía FAIR desde adentro. Lo que acaban de publicar es la versión que les dejan publicar. Pero con eso, es más que suficiente para decirles exactamente que es lo que está pasando. TRIBE v2 predice, vértice por vértice sobre la corteza cerebral, qué zonas activa cualquier video. Sin escáneres. Sin humanos. Subes el contenido, obtienes el mapa neural (activación emocional, supresión de razonamiento crítico, modulación prefrontal) antes de que el video lo vea un solo usuario. Ahora considera la posición de Meta: 1. Tiene años de datos de Reels sobre qué contenido retiene atención, genera enojo, provoca compartir. 2. Saben empíricamente qué funciona. TRIBE v2 les da el mecanismo causal de por qué funciona (a nivel de tejido cortical) Eso convierte correlación histórica en capacidad predictiva sobre contenido nuevo. 3. Internamente hay herramientas que se llaman Gatekeepers y Quick Promotions que sirven para inyectar contenido en el feed de poblaciones arbitrarias a escala. 4. Simulador de respuesta cerebral + conocimiento empírico de contenido efectivo + maquinaria de distribución selectiva. El pipeline está completo. Y luego está Thiel. Inversor y amigo personal de Zuck. Fundador de Palantir, cuyo negocio es análisis de poblaciones a escala para gobiernos e inteligencia. NO es descabellado observar que confluyen los incentivos de plataformas construidas por las mismas personas. La licencia CC BY-NC dice que Meta retiene los derechos comerciales del predictor de respuesta cerebral más preciso jamás construido. Y recuerda, esto es lo que decidieron hacer público.
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

Español
195
3K
12.3K
1.3M
Carles Lázaro Costa retweetledi
Durgan A. Nallar 🎲
Durgan A. Nallar 🎲@Durgan·
No sé quién es el autor, pero me pareció correcto: "Si Cthulhu puede ser invocado por humanos que están muy por debajo de él, ¿por qué los humanos no pueden ser invocados por hormigas? La respuesta es que deberían serlo. Bueno, si un montón de hormigas formaran un círculo en mi casa, ciertamente lo notaría, trataría de averiguar de dónde vienen todas, y posiblemente causaría una gran destrucción en ese sitio. Por eso saber y pronunciar correctamente el nombre verdadero es tan importante para el ritual. Imagina lo imposible que sería no ir a echar un vistazo si el círculo de hormigas empezara a cantar tu nombre. Y te dicen, no puedes irte porque dibujamos una línea hecha de pequeños cristales, ahora tienes que hacernos un favor. Y tú piensas, veamos a dónde va esto y dices: "Este... oh! sí, me tienes... ¿cuál es el favor?" y normalmente el favor es algo como, "mata a esta hormiga por nosotros" o "dame un montón de azúcar" y tú dices... ¿vale? y lo haces, porque por qué no, no es difícil para ti y vaya que esta va a ser una historia de la hostia para contar, estas putas hormigas cantando tu nombre y queriendo una cucharada de azúcar o lo que sea. Y A VECES te piden cosas que realmente no puedes hacer, una de ellas, ella dice, "amo a esta hormiga pero no me presta atención, hazme importante para ella" y tú dices... ¿eh? ¿cómo? Así que simplemente matas a todas las hormigas de la colonia excepto a estas dos, y ¡ta-da! ¡problema resuelto! y la primera hormiga esta como *susurro horrorizado* "¿Qué he hecho?"
Español
21
269
1.4K
82K
Carles Lázaro Costa retweetledi
Julius Master N♱N
Julius Master N♱N@JuliusMasteroso·
Vivimos en las ruinas de una civilización superior
Español
755
15.8K
64.3K
2.2M
Facuu
Facuu@Piponki_facu·
@Carleslc @crystalia414 Uffff es muy buena, cuando llegue a casa lo pruebo, lo q no veo en ningún lado es de donde saca los mangas, de curioso nomas
Español
1
0
1
34
🌸 Nerea 🌸
🌸 Nerea 🌸@crystalia414·
Por fin he encontrado una página web de mangas para descargar CON LAS TRADUCCIONES OFICIALES DE ESPAÑA 😭😭😭😭
Español
414
112
3.6K
144.1K
Facuu
Facuu@Piponki_facu·
@crystalia414 Los descarga como pdf? Sería un golazo para el Kindle
Español
2
0
2
3.1K
Carles Lázaro Costa retweetledi
meatball times
meatball times@meatballtimes·
software engineering is absolutely not being automated right now. I barely ever code any more but I have never been busier. now that coding is 80% automated, the limiting factor is my ability to design, comprehend, and safely change systems. it's insanely exciting
English
108
186
3.6K
104.1K
Carles Lázaro Costa retweetledi
Aiden Guo
Aiden Guo@aidenguoai·
This is why I worked on One Piece.
English
105
785
7.2K
418.8K
Carles Lázaro Costa retweetledi
Jen 🌮🌯
Jen 🌮🌯@jroebuck·
I mapped all the digital applications that give us dopamine hits through short term satisfaction and in theory help us recover from so much dopamine exposure.. AI is a replacement for social media and it does make work more enjoyable and addictive for many is us - for exactly that reason.. instant feedback loops bring us energy and motivation to create - it’s worse than short form video because it’s personalized.
Jen 🌮🌯 tweet media
English
3
8
50
6K
Carles Lázaro Costa retweetledi
Aakash Gupta
Aakash Gupta@aakashgupta·
This is a dopamine loop, and it’s one of the most powerful ones humans have ever encountered. Every time you prompt an AI and get a useful result back in seconds, your brain gets a hit. Variable-ratio reinforcement, same mechanism as slot machines, except the reward is real: actual output, actual progress, actual leverage on your ideas. Traditional work follows a delayed-reward structure. You write code for 6 hours, maybe it compiles, maybe you get feedback in a week. The gap between effort and reward is wide enough that motivation decays constantly. AI compresses that loop to seconds. Effort → reward → effort → reward. Your prefrontal cortex stays engaged because the next payoff is always one prompt away. This is why people describe it as “fun” when they’re actually working 14-hour days. The subjective experience of effort disappears when reward frequency is high enough. The “harder than ever” part is real too. When your bottleneck shifts from execution to imagination, you run out of excuses to stop. There’s no “waiting on the build” or “blocked by review.” Every idea you have can be tested immediately, which means your brain never gets a natural stopping point. People who thrive on this are selecting for a specific neurotype: high novelty-seeking, high conscientiousness, tolerance for rapid context-switching. That’s maybe 10-15% of the population. The other 85% will experience the same tools as overwhelming, not energizing. And that split is going to define the next decade of who captures value from AI and who gets displaced by it.
Nat Eliason@nateliason

Nearly every ambitious person I know who has dived into AI is working harder than ever, and longer hours than ever. Fascinating dynamic tbh. I have NEVER worked this hard, nor had this much fun with work.

English
207
487
4.6K
563.8K
Carles Lázaro Costa retweetledi
Eric S. Raymond
Eric S. Raymond@esrtweet·
If you are a software engineer "experiencing some degree of mental health crisis", now hear this, because I've been coding for 50 years since the days of punched cards and I have a salutary kick in your ass to deliver. Get over yourself. Every previous "programming is obsolete" panic has been a bust, and this one's going to be too. The fundamental problem of mismatch between the intentions in human minds and the specifications that a computer can interpret hasn't gone away just because now you can do a lot of your programming in natural language to an LLM. Systems are still complicated. This shit is still difficult. The need for people who specialize in bridging that gap isn't going to go away. As usual, the answer is: upskill yourself and adapt. If a crusty old fart like me can do it, you can too.
Tom Dale@tomdale

I don't know why this week became the tipping point, but nearly every software engineer I've talked to is experiencing some degree of mental health crisis.

English
657
1.8K
15.7K
1.6M
Carles Lázaro Costa retweetledi
World of Statistics
World of Statistics@stats_feed·
Neuroscientist study reveals that Gen Z has become the first generation to be less intelligent than its predecessor, the Millennials. For the first time in 100 years, young people are scoring lower than their parents on IQ tests and core skills like memory, reading, and focus. This is happening mainly in the US and Europe. The main cause appears to be excessive screen time and digital device use, particularly in schools and social settings. While some suggest Gen Z is just developing different skills, research shows actual declines in fundamental problem-solving abilities.
English
1.4K
4.5K
26.2K
5.5M
Carles Lázaro Costa retweetledi
Aakash Gupta
Aakash Gupta@aakashgupta·
This is one of the most important studies in sleep science. Van Dongen et al. ran the experiment that changed how we understand chronic sleep restriction. They had subjects sleep 4h, 6h, or 8h nightly for 14 days, testing cognitive performance every 2 hours. The 6h group’s reaction time deficits by day 14 matched subjects who had been awake for 24 hours straight. The 4h group? They performed like someone awake 48 hours. But here’s what makes this study terrifying. The Stanford Sleepiness Scale ratings in Panel B plateau after day 3-4. Subjects stopped feeling more tired even as their cognitive performance continued deteriorating through day 14. Your subjective experience of fatigue is a lagging indicator that eventually just… stops updating. This explains why chronic undersleeping feels sustainable. You’ve adapted to feeling tired. Your prefrontal cortex hasn’t adapted to being impaired. The PVT (Psychomotor Vigilance Task) in Panel A measures lapses in attention. These are the moments where you’re staring at a screen and your brain simply checks out for 500ms. Every additional day of 6h sleep adds more lapses. The curve never flattens. Panel C and D show working memory and processing speed. Same pattern: continuous degradation with no subjective awareness. The practical implications: If you’re sleeping 6h and think you’re functioning fine, you’ve lost the internal calibration to know you’re not. The subjects in this study would have told you they felt “okay” while performing like they’d pulled an all-nighter. For anyone doing cognitively demanding work, this means you cannot trust how you feel. You need to track objective markers: error rates, decision latency, problem-solving throughput. Sleep need is biological, not negotiable. Most adults require 7-9 hours, and the research shows no population-level adaptation to chronic restriction. “I only need 6 hours” is almost always “I’ve forgotten what baseline cognition feels like.“
Bailey Klemmensen@iiKlemm

Competitive gamers: 6h of sleep is a hidden nerf. 6h/night for 2 weeks → reaction time & attention decline to the equivalent of pulling MULTIPLE all-nighters. Worst part: subjective fatigue plateaus, so you stop noticing. If you're not getting 7+ hours/night, you're trolling.

English
231
4.1K
31.4K
4.1M
One Pace
One Pace@OnePaceProject·
One Pace vs. One Piece - Remember that time a filler character intruded on a canon scene? ✂️
English
69
121
5.2K
551.5K
Carles Lázaro Costa retweetledi
Google DeepMind
Google DeepMind@GoogleDeepMind·
Step inside Project Genie: our experimental research prototype that lets you create, edit, and explore virtual worlds. 🌎
English
984
4.3K
34.6K
13.4M
Carles Lázaro Costa
Carles Lázaro Costa@Carleslc·
@CarlosGarcs @franperez_co No conoces la historia de Elon Musk. Estás suponiendo que es un gestor cuando la realidad es que sí es un experto en cohetes. No puedes decir que es un simple gestor alguien que diseñó un motor de cohete exitoso para mejorar la eficiencia y el coste de los motores rusos.
Español
0
0
0
17
Carlos Garcés
Carlos Garcés@CarlosGarcs·
@franperez_co Curioso. Melon Musk puede decir que el responsable de una aerolínea es un contable, pero nosotros tenemos que suponer que Musk es un experto en cohetes cuando es un simple gestor.
Español
10
1
29
16K
Fran Perez
Fran Perez@franperez_co·
Los acontecimientos: 1. Elon empieza a meter Starlink en todas las aerolíneas. 2. Ryanair dice que no, porque “la antena aumenta el consumo de combustible”. 3. Otras aerolíneas muestran números: no aumenta, baja. 4. El CEO de Ryanair retruca: “ok, sube 2% y nadie va a pagar USD 1 extra por WiFi”. 5. Elon dice: “quizás estás desinformado” y publica los cálculos de un Boeing 737-800. 6. El CEO de Ryanair responde que Elon es un pelotudo y “no entiende la resistencia del aire” 🍿 7. Twitter explota: el CEO de la peor aerolínea del mundo sobrando al fundador de la mejor empresa aeroespacial de la historia 🍿🍿 8. Elon contesta que el pelotudo es él y pregunta si debería comprar Ryanair solo para echarlo y poner a alguien que se llame Ryan 🍿🍿🍿 9. Ryanair saca un comunicado oficial insultando a Elon 🍿🍿🍿🍿 Aviación low cost, pero drama premium.
Fran Perez tweet mediaFran Perez tweet mediaFran Perez tweet mediaFran Perez tweet media
Español
383
2.4K
28.9K
1.8M
Carles Lázaro Costa retweetledi
Jakes in Japan 🇯🇵
Jakes in Japan 🇯🇵@jakesonaplane_·
These greens and red tonight are insane
Jakes in Japan 🇯🇵 tweet mediaJakes in Japan 🇯🇵 tweet mediaJakes in Japan 🇯🇵 tweet mediaJakes in Japan 🇯🇵 tweet media
English
375
9K
53.9K
2.1M
Carles Lázaro Costa retweetledi
il Donaldo Trumpo
il Donaldo Trumpo@PapiTrumpo·
OK, WHO DID THIS???🤣🤣🤣
English
5.1K
41.7K
226.9K
12M
Carles Lázaro Costa retweetledi
One Piece Defender ◍
One Piece Defender ◍@OnePieceDefend·
One Piece season 1 ends after 1155 episodes. Here's one second from each episode until Gear 5.
English
128
6.1K
22.8K
701.8K