demis hassabis mass-won the nobel prize in chemistry in 2024 for alphaFold. what nobody reported is what he said at a private dinner in london three weeks later. "we are mass-building the most inefficient thinking machines in the history of engineering. and we are calling it progress." a neuroscientist at UCL who was at that dinner spent the next 18 months turning that sentence into a 37-page quantitative proof. the paper was submitted to nature in august 2026. it was rejected. not because the science was wrong. because the implications were "too commercially sensitive to publish without further review." the paper leaked on a closed neuroscience slack on october 1st. Working Note: The Efficiency Gap Biological vs Silicon Intelligence - A Thermodynamic Analysis August 2026 here's what the paper proves - the human brain runs at 20 watts. the same power as a dim lightbulb. it processes approximately 10^16 operations per second. GPT-4 inference on an NVIDIA H100 cluster: 3,200,000 watts. it processes approximately 10^15 operations per second. the brain is 10x faster on 160,000x less energy. the efficiency ratio: brain = 5 × 10^14 ops/watt. H100 cluster = 3 × 10^8 ops/watt. the brain is 1,600,000x more energy-efficient than the best AI hardware on earth. now watch the historical curve [Fig. 1] - 2012 (AlexNet era): AI was 100,000,000x less efficient than the brain. 2017 (transformer era): 10,000,000x less efficient. 2023 (GPT-4 era): 1,600,000x less efficient. 2026 (current): 1,200,000x less efficient. at the current rate of improvement - 22% efficiency gain per year - AI hardware will match the brain's efficiency in 2089. sixty-three years from now. jensen huang's entire NVIDIA roadmap assumes exponential improvement. the paper proves the improvement is logarithmic. it's slowing down, not speeding up. the reason is thermodynamic. the landauer limit. rolf landauer proved in 1961 that erasing one bit of information requires a minimum energy of kT ln(2). at room temperature: 2.85 × 10^-21 joules. silicon transistors currently operate at approximately 1,000x above the landauer limit. every hardware generation closes that gap by roughly 15%. but the closer you get to the limit, the harder each improvement becomes. this is why moore's law died. the brain operates at 10x above the landauer limit. not 1,000x. ten. 3.8 billion years of evolution optimized the brain to within one order of magnitude of the physical limit of computation. NVIDIA has had 31 years and is still three orders of magnitude away. the paper asks one question in the conclusion: "what architectural principle allows biological neural networks to operate within 10x of the thermodynamic floor while artificial networks remain at 1,000x?" the answer: the brain doesn't separate memory and processing. every neuron is both storage and computation simultaneously. every synapse is both a wire and a processor. every computer built since von neumann's 1945 architecture separates memory from compute. data moves back and forth on a bus. the bus is the bottleneck. the bus is where the energy goes. NVIDIA's entire product line - every GPU, every tensor core, every NVLink - is an increasingly elaborate attempt to make the bus faster. not to eliminate it. faster. the brain has no bus. it never needed one. the paper's final line: "the companies that build the first commercially viable non-von-neumann inference chip will make the GPU obsolete within 5 years. the physics demands it." intel abandoned their neuromorphic chip program (loihi) in 2023. IBM shelved TrueNorth. the two companies closest to solving the problem quit. the $3.4 trillion NVIDIA valuation is a bet that the bus can be made fast enough forever. landauer proved it can't. 65 years ago. save this

Hi. I am a wheelchair seller. Buy my wheelchair. What's that?? You don't NEED a wheelchair?? No worries, this here is Boris. He will now break both your legs and for good measure, sever your lower spine so you become a paraplegic. See?! I was right! You need my wheelchair!

Claude Code just shipped mods... I made one that connects you to a multiplayer Doom server while Claude is busy. Every other player in the match is a person waiting for their Claude to finish.

UCLA researchers built an image generation model that runs on “light” instead of GPUs. And it matches a 1.07 billion parameter diffusion model.. and generates each image in less than 1 nanosecond with basically zero computing power. It’s called “Optical Generative Model” Instead of pushing electrons through billions of transistors, it uses the physics of light to draw pictures. A digital encoder creates a noise pattern and imprints it onto a laser. The light then propagates physically through custom optical layers. The light itself does the calculating. There is no silicon bottleneck. It processes all-optically. And it uses almost zero compute power during the actual generation. It successfully created complex images, human faces, and Van Gogh-style art with the exact same quality as traditional AI. But the energy cost wasn't just reduced. It was practically eliminated. If this scales, the carbon footprint and computing limits of modern AI disappear overnight. We’ve spent years trying to optimize code to make models faster and cheaper. But the next massive leap in AI isn't going to be software.

Lazy work used to mean too little output. Now, with AI, it often means too much and more work for everyone else. @tobi call it "slop grenades." A "Slop Grenade" is when you let AI produce the work and pass it on without adding any value (including checking it). Someone else has to wade through it, catch the mistakes, and clean up the mess. You save time and look productive but someone else pays for it.

Google just won a bankruptcy auction for Spirit Airlines’ internal corporate data for $10 million and plans to use it to improve its products and train AI models. This isn't passenger profiles or credit-card data. Google is buying something potentially much more useful for enterprise AI: years of records showing how a real company actually operated. The dataset reportedly includes: • ~100 million employee emails • ~500 million Microsoft Teams messages and collaboration records • ~30 million lines of software code, plus development metadata, models, and algorithms • Spreadsheets, calendars, presentations, and project-management documents • Revenue, pricing, marketing, and financial databases • Aircraft operations data • Employee productivity records • Audit and fraud-related information • HR and strategy materials Google beat AI data company Mercor, which offered $7.5 million. The bankruptcy filing formally names Google as the successful bidder at $10M and Mercor as the alternate bidder at $7.5M

An AI just read a stranger's mind through their skull, no surgery, no implants, and got roughly 45% of the meaning right on the first try. On a person it had never seen before. That's the tech an OpenAI researcher just quit her job for. Two weeks ago she walked out of the hottest company on earth to join Conduit, a startup collecting brain recordings in a San Francisco basement. Here's what happens in that basement. Paid volunteers put on four pound headsets covered in sensors and chat with an AI for two hours while it listens to their neurons fire. The signal is so faint the team shut off wall power to the entire building and runs on battery packs, because the hum from an ordinary outlet drowns out a human thought. They've collected 10,000 hours of brains so far. The largest dataset of its kind ever assembled. And here's the part that should make your hands sweat. Whisper needed 680,000 hours of audio to master speech. The moment someone scales brain data the same way, the decoding stops being 45% right. She says we're in the GPT-2 era of mind reading. GPT-2 was a joke in 2019. Five years later half the internet was written by its grandchildren. Her timeline says developers wear thought-streaming headbands by 2027. By 2035, AI stops feeling like a tool and starts feeling like a limb. Your inner monologue has been the one place nothing could touch for 300,000 years of human history. That run is ending in a basement in San Francisco, two hours at a time.

Palantir CEO Alex Karp on the Wealth Tax:  "It just punishes." "It starts with the billionaires." "Every single person at this table is going to be paying a wealth tax only to punish us." "And the reason for it is—these models have been completely, irresponsibly oversold." " And the sell is, 'It's dangerous for everyone—which is why I can give it to all your adversaries, but I can't give it to the Department of War." "Or, I can't safely give it to an enterprise in this country without being certain that the alpha of that business could transfer to this model tomorrow." "At every single enterprise I deal with, these people are livid." "They're like, 'I'm paying for tokens that create no value. These people are stealing the weights and alpha of my business, and they're creating a wealth tax that does not help the poor." Via @CNBC