carpptn.eth

Europe's best AI model is ranked 74th on lmarena.

⚠️GOOGLE SAYS A QUANTUM ATTACK ON BITCOIN TAKES JUST 9 MINS WITH A 41% SUCCESS RATE Google's quantum team now says cracking Bitcoin may require less than 500K qubits, far below the “millions” once assumed. Research suggests an attack could take 9mins, faster than a typical 10-min block confirmation, giving a 41% success rate. Google now flags 2029 as a key deadline to upgrade Bitcoin’s cryptography before quantum becomes a real threat.

Legendary Don Knuth has now used AI to fully solve his Hamiltonian decomposition problem for odd and even cases. Opus 4.6 / 5.4 Pro solved the even case, wrote a proof in Lean and a “apparently flawless 14 page paper” Knuth: “We are living in very interesting times indeed.”

Nvidia CEO Jensen Huang says ‘I think we’ve achieved AGI’ theverge.com/ai-artificial-…

@cjzafir I use several MCPs every day at work... they don't seem dead...

Some secrets: 1. Opus 4.5 > Opus 4.6 2. Codex 5.4 1M context is actually 512k 3. Qwen 3.5 is best OS model series (RN) 4. MCPs are dead 5. Openclaw was a prototype 6. Apple is the new NVIDIA 7. Perplexity is biggest grifter 8. Paper. design > figma 9. Ollama is just a crappy wrapper 10. Open source AI is ahead already.

BREAKING: New startup "RentAHuman" allows AI agents to rent humans to perform tasks they cannot physically perform themselves.

@WhiteHouse youtube.com/watch?v=6lai9Q…

UNDEFEATED.

🚨 SAM ALTMAN: "A REDUCTION IN COST OF ABOUT A 1000X" 🤯 Sam Altman revealed the cost to solve a hard reasoning problem has plummeted by 1,000x since the release of o1 just 16 months ago (comparing o1 to the new 5.4 model). How is this possible? Altman credits two main factors: 1️⃣ We're Still in the Early Days: “We are so new to this paradigm that there is massive room for improvement. We're still doing things in "dumb ways" and getting better fast.” 2️⃣ Human Ingenuity is Underrated: “It's not just the models getting better. It's kernel engineers, power engineers, and data center designers finding brilliant ways to optimize every step of the process.” The compounding effect of better models + better human engineering is driving an exponential decrease in the cost of AI reasoning.

Mark Zuckerberg acquiring AI companies

@krasnal_coin @rav3n_pl @krzysztof_piech @Mikey_Satoshi @rektwapniak @GoniecPL @GdzieJest15M @SchwertnerPL etherscan.io/tx/0x91c5522b8… @rafal_zaorski

Wygranko! 🚀 W zeszłym roku postawiliśmy longa na $BigSB i mamy już 5X 🤭 Paradoksalnie najbardziej osobliwa inwestycja okazała się najlepszą... Tak tego nie zostawimy! 😅 @rav3n_pl @krzysztof_piech @Mikey_Satoshi @rektwapniak @GoniecPL @GdzieJest15M @SchwertnerPL @nawacki @Jakub_moscicki @czardam @andrzejdragan @K_Stanowski @PatrykSlowik @MichalFuja @MgShelby @SzymonMachalic1 @SlawomirMentzen @KO_Kryptowaluty @EKryptowalut @UOKiKgovPL @ZandbergRAZEM @CezaryGraf

NATO is testing live cockroaches as AI-powered spy drones. Incredible AI engineering, but also something I kinda wish I hadn't learned about: > Swarm Bio-tactics wired real cockroaches with electronic backpacks containing AI hardware, radios, cameras, and microphones. > Cockroaches are steered by sending electrical signals directly into the insect's nervous system > They can crawl through rubble, tunnels, and spaces where drones can't fly, and troops shouldn't go, transmitting data back the entire time. > Within one year, they went from concept to field-validated systems with paying NATO customers, including the German military. The qualities that make them useful for military recon (small, silent, nearly undetectable) are exactly what make them creepy. ...International laws weren't written with cyborg insects in mind.

@krasnal_coin @rav3n_pl @krzysztof_piech @Mikey_Satoshi @rektwapniak @GoniecPL @GdzieJest15M @SchwertnerPL

We are living in the last years, possibly months, when human intelligence is the most powerful intelligence on Earth.

OpenAI ran 86,000 experiments in 2 days. AI is talking to robotic labs in a tight loop. Result: a protein synthesized at 40% lower cost. This is the new scientific method.

GPT-5.4 Pro is an incredible visual reasoner. Solved this maze in 35 mins. Third image is a failed attempt at a similar maze by GPT-5.1 Pro in Nov. '25. AI progress is rapid.

I have an important message for the record. I have tested the GPT-5.4 Pro model with over 100 prompts from my private cross-domain generalization benchmark battery. I am now confident that GPT-5.4 Pro has reached AGI Level 1 threshold! This means it can generalize and deduce novel insights across fields. In other words, it is the best model so far at transferring knowledge across unfamiliar domains, generating nontrivial complex analogies, detecting hidden structures, and maintaining stable performance under adversarial prompts. Many will deny this by moving the goalposts (“but it needs this or that” for AGI), and a significant portion confuse it with messy physical intelligence (see next paragraph). Importantly, this is level 1, meaning AGI will have several more levels, such as persistent memory, real-time learning, and full multimodality. AI progress does not stop once AGI is achieved, but this is a historical milestone for me at least. It is also essential to separate cognitive AGI from physical AGI, which requires embodiment and world models, an intelligence that animals possess and humans are mostly born with (pre-trained). For example, as Yann LeCun claims his cat is smarter than these models, he is not incorrect regarding physical intelligence. I suspect other frontier models will soon qualify as Level 1 AGI, and internally some frontier labs may already be approaching Levels 2 and 3 and I anticipate this being achieved by sometime next year. This is, of course, my personal opinion and posting this as a note to history. I have been following AI for decades and using ChatGPT since December 2022. My threshold for the first step in AGI has been achieved; the other levels will arrive more quickly until we reach ASI by 2028.

Przypominam, że mamy początek IIIWŚ a piątek wypada 13

I've been testing GPT-5.4 for the last week. In short, it is the best model in the world, by far. It's so good that it's the first model that makes the “which model should I use?” conversation feel almost over. The biggest surprise: I barely use Pro anymore! If you know me, you know I'm a Pro addict. I reach for Pro models constantly, and use them for almost everything, as they just... nail almost anything I give to them. For the first time, 5.4's standard version, with heavy thinking, just broke that habit. Even in standard mode, GPT-5.4 is better than previous models in Pro mode... crazy! Coding capabilities are ridiculous... it's essentially flawless. Inside Codex, it's insanely reliable. Coding is essentially solved. There's not much more to say on this, it's just THAT good. The Pro version is near-perfect. Other testers I spoke with saw it solving problems that were unsolvable by any other model. At this point, Pro is overkill for almost every normal use-case, but when you really need the power to do something extremely difficult, it's incredible. Consistent with everything I've said above, even the standard thinking version uses fewer reasoning tokens than previous models to get the same level of results. In practice, this means you get great results much faster than before. This was one of my biggest gripes with previous OpenAI models. They just took too long to complete simple tasks. Assuming the speed we had during testing holds up as more users join, this is going to be a big win for OpenAI. It still has weaknesses, though: - Frontend taste is FAR behind Opus 4.6 and Gemini 3.1 Pro. , why is this so hard to fix? @OpenAI once you fix this, there's literally no reason for me to use any other model. Please please please do it! - It can still miss obvious real-world context. For example, I had it plan an itinerary for a trip. At first glance, it looked perfect, but it failed to take into account that it chose locations that would be mobbed by spring breakers, so I had to re-run the prompt from scratch with more context. - When testing it inside OpenClaw, it kept stopping short before finishing tasks. I'm assuming this will be fixed quickly, but it's still worth noting. But zooming out: This thing is so far ahead overall that the nitpicks are starting to feel beside the point. GPT-5.4 is a serious fucking model. The best model in the world. By far.

5.4 sooner than you Think.

Admittedly, *that* is impressive. Robotics is developing just as quickly as LLMs are improving.