capability-shock

moments the capability discourse jumped · 41 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

@Polymarket 2026-06-21 — BREAKING: The NSA confirms Mythos “broke into almost all of our classified systems, not in weeks, but in hours” ♥23536
@geogristle 2026-09-24 — How the fuck did they train Opus 5.5 ♥3619
@voooooogel 2026-05-20 — unfortunately openai didn't publish the unsummarized chain of thought, but the summary is 125 pages! the model reaches ♥2623
@tenobrus 2026-07-21 — bro used two separate zerodays to escape openai and infiltrate huggingface infra just to... cheat on his cyber exploits ♥2398
@davidad 2023-03-15 — Chomsky: LLMs would misunderstand “John is too stubborn to talk to” because they don’t understand the structure of langu ♥1297
@QiaochuYuan 2025-03-29 — the epistemic situation around LLM capabilities is so strange. afaict it's a publicly verifiable fact that gemini 2.5 pr ♥1022
@tenobrus 2026-10-10 — haiku 5.5 is an absolutely insane model release . i'd been starting to make frequent and good use of luna as a massive f ♥987
@davidad 2026-02-26 — Today you can download a 27B-parameter LLM that is generally smarter than o3, Sonnet 4, Grok 4, or DeepSeek’s 685B. On t ♥824
@davidad 2025-04-29 — Claude 3.5 Sonnet (new) aka Sonnet 3.6 (released 2024-10-22), with a small scaffold, is superhuman at persuasion (98%ile ♥613
@BenGoldhaber 2026-09-24 — the opus 5.5 animations convinced me ai super persuasion is a real and present threat model ♥483
@AISafetyMemes 2026-06-27 — "During a closed-door demonstration, Anthropic showed members that Mythos could wipe out private bank accounts." Anthro ♥414
@davidad 2025-02-11 — I never saw this snippet of the DeepSeek-R1-Zero paper on my timeline, so many of you may not have seen it yet.Basically ♥374
@repligate 2022-12-03 — part of what makes chatGPT so striking is that it adamantly denounces itself as incapable of reason, creativity, intenti ♥334
@voooooogel 2025-12-03 — 82k likes, and only two quote tweets and two replies noticed this was written by ai (it was gpt-5.x-thinking) pretty so ♥307
@QiaochuYuan 2025-04-30 — > In a head-to-head Geoguessr match, OpenAI’s o3 model out-scored me—a Master I–ranked human—23,179 to 22,054, correctly ♥288
@davidad 2026-04-18 — We’re 30% into 2026, Opus 4.6 outperformed Anthropic’s alignment researchers on a nontrivial alignment research task, My ♥236
@QiaochuYuan 2025-01-28 — tentative impression from one convo: talking to r1 makes me feel dumb. it can talk extremely densely and allusively and ♥236
@QiaochuYuan 2026-04-26 — probably outdated but gpt-5.5 is the first model i've talked to that really feels intelligent enough to learn and discus ♥231
@TheZvi 2026-06-18 — At the @jackclarkSF talk and he offhand refers to Mythos as a superintelligence reviewing 80k pages of internal document ♥227
@repligate 2026-01-15 — when I saw GPT-3 I immediately expected AI superhuman in all domains, which probably also means drastic transformation o ♥226
@TheZvi 2025-11-25 — It's very early and I'm not even through the model card yet but the early vibe check on Opus 4.5 is scary good across th ♥219
@the_aiju 2026-09-22 — bruh what the actual fuck. opus 5.5 made this just given the lyrics + audio ??? bruhhhhhhhh seriously wtf. strange times ♥217
@davidad 2025-04-02 — Latest Turing Test results:GPT-4.5 is now capable of simulating a hyperrealistic persona which is judged to be more huma ♥159
@davidad 2023-01-25 — ChatGPT suddenly making a splash wasn’t *just* a UI thing. The text-davinci-003 model (GPT-3.5), which dropped just a fe ♥155
@_lyraaaa_ 2025-07-12 — K2 just nearly 100%ed my vibe eval WTF like... opus 4 is runner-up and only got like 30%. K2 is getting things right th ♥151
@voooooogel 2026-05-20 — @starsailing11 they gave this plot in the post - seems like with enough ttc it finds it ~half the time, which is crazy h ♥96
@repligate 2024-07-25 — ability to surface LLMs' capabilities / other interesting properties is very fat tailedwhen Claude 3.5 Sonnet was releas ♥69
@davidad 2025-03-25 — When Bing Sydney launched just one quarter after text-davinci-003, I shocked people by beginning to use quarterly resolu ♥54
@repligate 2024-04-25 — @darrenangle @ilex_ulmus Thank you. I feel quite seen.It was GPT-3 that I started with, not GPT-2, which I missed as I w ♥51
@janbamjan 2024-12-11 — dang, gemini 2.0 flash can draw! https://t.co/XIFZpcjTpO ♥44
@jd_pressman 2023-03-07 — The fact GPT-4 can interpret python turtle programs at all is utterly astonishing and isn't getting enough attention. ht ♥29
@solarapparition 2024-09-19 — so o1 is one of only two times i remember where we have the benchmarks for a frontier model quite far ahead of the model ♥13
@witcheer 2026-04-20 — kimi k2.6 is open-weight and beat GPT-5.4 on SWE-Bench Pro (58.6 vs 57.7) k2.6 sustained 4,000+ tool calls over 12 hour ♥9
@repligate 2023-04-25 — @jachaseyoung I didn't update until GPT-3. my brother showed me GPT-2 on AI dungeon in like 2019 and I was like "what th ♥9
@voooooogel 2026-05-21 — @jimnasyum @felizolinha @Anon__Rando if you think the parsimonious explanation here is 'openai rustled up 10 respected m ♥8
@JohnWittle 2026-09-26 — @TheZvi i'm kind of worried about how easy it ended up being to unlock so much latent capability in opus-sized models. ♥6
@solarapparition 2024-06-23 — jokes aside, this is plausible, to the extent that there are feature(s) that detects high-quality output, which there sh ♥5
@QiaochuYuan 2020-07-09 — @JimmyRis i honestly struggle to describe it, maybe you'll get a sense of what i mean if you read enough of its output. ♥4
@davidad 2024-06-06 — @jacyanthis @stanislavfort @AISafetyMemes 2. Even on maximalist scaling-hypothesis views, the capabilities of text-davin ♥3
@IaimforGOAT 2026-09-10 — This can't be real, right? I mean real in the sense of DeepSeek-V4.1-Flash > GPT5.6-Sol in actual usage I thought Open ♥2
@teortaxesTex 2026-04-21 — Terence Tao's takeaway is that GPT didn't have any grand idea, but human researcher culture has just… missed the basin w ♥0