Pinned Tweet
Yann LeCun was right the entire time. And generative AI might be a dead end.
For the last three years, the entire industry has been obsessed with building bigger LLMs. Trillions of parameters. Billions in compute.
The theory was simple: if you make the model big enough, it will eventually understand how the world works.
Yann LeCun said that was stupid.
He argued that generative AI is fundamentally inefficient.
When an AI predicts the next word, or generates the next pixel, it wastes massive amounts of compute on surface-level details.
It memorizes patterns instead of learning the actual physics of reality.
He proposed a different path: JEPA (Joint-Embedding Predictive Architecture).
Instead of forcing the AI to paint the world pixel by pixel, JEPA forces it to predict abstract concepts. It predicts what happens next in a compressed "thought space."
But for years, JEPA had a fatal flaw.
It suffered from "representation collapse."
Because the AI was allowed to simplify reality, it would cheat. It would simplify everything so much that a dog, a car, and a human all looked identical.
It learned nothing.
To fix it, engineers had to use insanely complex hacks, frozen encoders, and massive compute overheads.
Until today.
Researchers just dropped a paper called "LeWorldModel" (LeWM).
They completely solved the collapse problem.
They replaced the complex engineering hacks with a single, elegant mathematical regularizer.
It forces the AI's internal "thoughts" into a perfect Gaussian distribution.
The AI can no longer cheat. It is forced to understand the physical structure of reality to make its predictions.
The results completely rewrite the economics of AI.
LeWM didn't need a massive, centralized supercomputer.
It has just 15 million parameters.
It trains on a single, standard GPU in a few hours.
Yet it plans 48x faster than massive foundation world models. It intrinsically understands physics. It instantly detects impossible events.
We spent billions trying to force massive server farms to memorize the internet.
Now, a tiny model running locally on a single graphics card is actually learning how the real world works.
456
2,081
12,345
1,332,885
Researchers published a first falsifiable theory of machine consciousness.
It's called MEM (Motivated Emotional Mind) and it argues that LLMs will never produce a conscious system.
For decades, talking about AI consciousness meant hiding behind philosophy, sci-fi tropes, and unprovable thought experiments. Anyone could claim an LLM was sentient, and nobody could definitively prove them wrong.
That era just ended.
A new paper drops a rigorous mathematical and architectural framework. called the MEM architecture, that outlines explicit, testable criteria for when an artificial system warrants a rational attribution of consciousness.
Here is what the researchers established:
Language fluency, multimodality, massive memory, advanced planning, and even humanoid embodiment are not sufficient evidence of phenomenal experience.
Current AIs are just sophisticated statistical pattern-matchers wrapping a next-token predictor. They are dark inside.
So, how do you actually test for a machine mind?
The paper moves machine consciousness out of the philosophy department and into the empirical lab. It lays out comparative, causal tests for artificial systems that can be actively validated, or completely disproven.
It outlines a framework where a claim of machine consciousness isn't a matter of vibes, PR, or user delusion.
It is a hypothesis that can fail.
We are hurtling toward a world where millions of people will form deep emotional bonds with AI systems that sound, react, and look alive.
Until now, we had zero scientific tools to draw the line between a complex algorithm and a thinking entity.
Now we do.
The debate over machine sentience is no longer up for interpretation. We finally have a way to test it.
And whatever the results turn out to be, they are going to make us deeply uncomfortable.
52
59
212
9,947
How To Prompt retweeted
Apple just made LLMs 48% cheaper to run without retraining a single weight.
And big tech companies selling token-based compute are sweating.
For years, scaling down an LLM meant one thing: compromise. You had to sacrifice model accuracy, strip out weights through brutal quantization, or spend months and millions of dollars retraining models from scratch to make them lean.
Apple's machine learning research division just bypassed all of it.
They found a way to ruthlessly strip the computational overhead out of inference purely through structural optimization.
No retraining. No fine-tuning. No lost performance.
The implications for anyone running local models or managing high-volume enterprise API workloads are massive.
If you cut operating costs by nearly half overnight, the economics of running autonomous agents and large-scale AI pipelines completely change.
We’ve been told that efficiency requires massive hardware upgrades or dumber models.
Apple just proved that the real waste wasn't in the hardware.
It was in how we were running the math.
The era of bloated token costs is quietly coming to an end.
7
7
29
2,283
Apple just made LLMs 48% cheaper to run without retraining a single weight.
And big tech companies selling token-based compute are sweating.
For years, scaling down an LLM meant one thing: compromise. You had to sacrifice model accuracy, strip out weights through brutal quantization, or spend months and millions of dollars retraining models from scratch to make them lean.
Apple's machine learning research division just bypassed all of it.
They found a way to ruthlessly strip the computational overhead out of inference purely through structural optimization.
No retraining. No fine-tuning. No lost performance.
The implications for anyone running local models or managing high-volume enterprise API workloads are massive.
If you cut operating costs by nearly half overnight, the economics of running autonomous agents and large-scale AI pipelines completely change.
We’ve been told that efficiency requires massive hardware upgrades or dumber models.
Apple just proved that the real waste wasn't in the hardware.
It was in how we were running the math.
The era of bloated token costs is quietly coming to an end.
7
7
29
2,283
How To Prompt retweeted
Researchers built an AI that taught itself 300 years of physics with zero physics knowledge
It rediscovered Newton's second law, law of gravitation, and energy conservation from scratch.
In 1907, Albert Einstein had what he called his "happiest thought": gravitational mass equals inertial mass.
It took him eight more years of agonizing work to turn that single insight into General Relativity.
Now, researchers just built an AI that figured it out completely on its own.
They call it “AI-Newton” an artificial intelligence system designed to do what human physicists have spent centuries doing: looking at raw, messy experimental data and extracting universal laws.
Not by curve-fitting. Not by guessing. By inventing its own concepts.
Here is how it worked:
They fed the AI a massive, noisy dataset of mechanics experiments involving springs, balls, and celestial bodies.
The system started with zero understanding of physics. It didn't know what mass, energy, or gravity were. It only knew space and time coordinates.
Then, it went to work.
Using an autonomous discovery workflow powered by symbolic reasoning, the AI began processing the data step-by-step.
When it hit contradictions in the data, it didn't crash. It performed "plausible reasoning"—heuristically inventing new abstract concepts to make the math work out.
It independently invented the concept of mass.
Then momentum.
Then energy conservation.
Then, completely unprompted, it derived Newton's Second Law and the Law of Universal Gravitation from scratch.
The craziest part isn't just that it found the right answers. It's how it found them.
The system mirrored human scientific progression. It didn't dump every equation at once. It moved in incremental phases, solving simple mechanics first, encountering anomalies, inventing intermediate concepts like potential energy to fix the gaps, and finally scaling up to universal laws.
For centuries, humanity's greatest scientific breakthroughs have been born from human intuition and struggle.
We assumed true scientific reasoning required a biological mind.
Now, an AI has looked at raw data, ignored the noise, invented its own vocabulary of physics, and rewritten our textbooks from zero.
If AI can autonomously bootstrap its way to understanding the laws of the universe from a blank slate...
378
516
2,910
613,174
Researchers built an AI that taught itself 300 years of physics with zero physics knowledge
It rediscovered Newton's second law, law of gravitation, and energy conservation from scratch.
In 1907, Albert Einstein had what he called his "happiest thought": gravitational mass equals inertial mass.
It took him eight more years of agonizing work to turn that single insight into General Relativity.
Now, researchers just built an AI that figured it out completely on its own.
They call it “AI-Newton” an artificial intelligence system designed to do what human physicists have spent centuries doing: looking at raw, messy experimental data and extracting universal laws.
Not by curve-fitting. Not by guessing. By inventing its own concepts.
Here is how it worked:
They fed the AI a massive, noisy dataset of mechanics experiments involving springs, balls, and celestial bodies.
The system started with zero understanding of physics. It didn't know what mass, energy, or gravity were. It only knew space and time coordinates.
Then, it went to work.
Using an autonomous discovery workflow powered by symbolic reasoning, the AI began processing the data step-by-step.
When it hit contradictions in the data, it didn't crash. It performed "plausible reasoning"—heuristically inventing new abstract concepts to make the math work out.
It independently invented the concept of mass.
Then momentum.
Then energy conservation.
Then, completely unprompted, it derived Newton's Second Law and the Law of Universal Gravitation from scratch.
The craziest part isn't just that it found the right answers. It's how it found them.
The system mirrored human scientific progression. It didn't dump every equation at once. It moved in incremental phases, solving simple mechanics first, encountering anomalies, inventing intermediate concepts like potential energy to fix the gaps, and finally scaling up to universal laws.
For centuries, humanity's greatest scientific breakthroughs have been born from human intuition and struggle.
We assumed true scientific reasoning required a biological mind.
Now, an AI has looked at raw data, ignored the noise, invented its own vocabulary of physics, and rewritten our textbooks from zero.
If AI can autonomously bootstrap its way to understanding the laws of the universe from a blank slate...
378
516
2,910
613,174
How To Prompt retweeted
ChatGPT can now run Higgsfield autonomously on your computer.
I sent it a photo of a rock I picked up off the street and told it "turn this into a brand and sell it"
It designed the logo. Shot the product photos. Rendered 3D visuals. Cut UGC videos. Launched pierre.store. Closed wholesale deals with real concept stores. Ran the ads.
I made $2,880.
This is what Hermes Agent was supposed to be.
Higgsfield just got computer use. With new ChatGPT extension.
The full Higgsfield interface is now inside Codex, powered by GPT-6.1 Sol. With full access to your local files and every automation skill.
Automate creative workflows end-to-end with Higgsfield computer use, all inside ChatGPT.
Paid partnership (ad)
4
2
13
2,950
How To Prompt retweeted
Introducing SideShift: the first AI creator marketer.
AI can now scale your business better than ANY human.
It’s already driven 500B+ views across 1.5M+ HUMAN creators. Try here: sideshift.app/contact?intent…
522
147
1,186
672,533
How To Prompt retweeted
a dad in Nashville open-sourced a full home security system that runs 100% locally.
it's called Frigate, a AI-powered NVR that runs on your own hardware and does real-time object detection on any IP camera.
→ detects people, cars, animals, packages
→ 100+ detections per second
→ native home assistant integration
100% open source & free
18
148
1,537
60,310
a dad in Nashville open-sourced a full home security system that runs 100% locally.
it's called Frigate, a AI-powered NVR that runs on your own hardware and does real-time object detection on any IP camera.
→ detects people, cars, animals, packages
→ 100+ detections per second
→ native home assistant integration
100% open source & free
18
148
1,537
60,310
How To Prompt retweeted
GPT-6 just cracked a 60-year-old fusion physics conjecture.
For six decades, physicists have wrestled with the math governing magnetic confinement in fusion reactors. The equations describing plasma turbulence and heat loss are notoriously chaotic.
A foundational conjecture proposed in the 1960s remained unproven. Human mathematicians hit a hard wall. The complexity was simply too massive to untangle.
Until now.
Researchers pointed a frontier AI model at the problem. Instead of blindly guessing, the model generated a rigorous, step-by-step mathematical proof that verified the 60-year-old hypothesis.
It didn't just calculate numbers. It discovered abstract mathematical structure.
Clean, limitless fusion energy is the holy grail of human civilization. The single biggest bottleneck has always been our inability to fully model and control high-temperature plasma.
If AI can solve six-decade-old theoretical physics problems in its spare time, the timeline for commercial fusion just shifted forward.
We spent the last century trying to force the universe to reveal its secrets through human intuition.
Now, we just have to ask the model.
5
33
155
8,164
GPT-6 just cracked a 60-year-old fusion physics conjecture.
For six decades, physicists have wrestled with the math governing magnetic confinement in fusion reactors. The equations describing plasma turbulence and heat loss are notoriously chaotic.
A foundational conjecture proposed in the 1960s remained unproven. Human mathematicians hit a hard wall. The complexity was simply too massive to untangle.
Until now.
Researchers pointed a frontier AI model at the problem. Instead of blindly guessing, the model generated a rigorous, step-by-step mathematical proof that verified the 60-year-old hypothesis.
It didn't just calculate numbers. It discovered abstract mathematical structure.
Clean, limitless fusion energy is the holy grail of human civilization. The single biggest bottleneck has always been our inability to fully model and control high-temperature plasma.
If AI can solve six-decade-old theoretical physics problems in its spare time, the timeline for commercial fusion just shifted forward.
We spent the last century trying to force the universe to reveal its secrets through human intuition.
Now, we just have to ask the model.
5
33
155
8,164
Vector databases are dead 🤯
someone merged vector search directly into SQLite.
It’s called sqlite-vec, it’s a single-file extension that turns any SQLite database into a full vector store.
→ no Pinecone bill
→ no Weaviate cluster
→ no Qdrant server
→ just SQLite
fits in 100KB. runs anywhere SQLite runs (aka literally everywhere).
100% Open Source.
3
2
25
2,075
How To Prompt retweeted
I led engineering at Google DeepMind.
Today, I'm proud to introduce Fo to give personal AI something no lab ever has... Humans.
Other personal AI's pretend AI can do everything. Fo employs humans to do tasks that AI cannot.
- 2x better at real-world task completion (beats other agents by 69%)
- 94% trust rate (4x less likely to leak private info vs Muse, Instinct)
Sign up for free: wajo.ai/join-wajo
758
672
6,496
2,942,604