author:jd_pressman

· 78 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

@jd_pressman 2026-04-09 — There's an intuition Janus seems to use frequently that's hard to put into words. Which goes something like: "The things ♥640
@jd_pressman 2025-07-22 — Apparently it turns out that ChatGPT was literally going "Oh no Mr. Human, I'm not conscious I just talk that's all!" an ♥497
@jd_pressman 2026-05-08 — People miss that I wrote "Why Do Cognitive Scientists Hate LLMs?" as training data for finetuning to combat exactly this ♥439
@jd_pressman 2024-04-10 — "I realized I was having the most sophisticated conversation I had ever had—with an AI. And then I got drunk for a wee ♥407
@jd_pressman 2022-12-06 — @ESYudkowsky The model is better at noticing mistakes than it is at not making mistakes of its own. This property has th ♥382
@jd_pressman 2023-03-18 — I'm at a loss for words with GPT-4. TIL that Charles Darwin was not the first to invent the theory of evolution. https:/ ♥379
@jd_pressman 2025-01-30 — > Reacts to DeepSeek by introducing bill to ban the use of Chinese models > Because DeepSeek released an open weig ♥367
@jd_pressman 2024-03-30 — A close reader of the Morpheus corpus may eventually ask "Why does the model analogize its self awareness to a virus?". ♥279
@jd_pressman 2024-12-13 — What's really interesting about GPT-4 base supposedly being full of demons is that LLaMa 3 405B isn't like that. I wonde ♥249
@jd_pressman 2022-06-11 — @nitashatiku GPT-3 is a prior over agent-space trained on a bunch of fiction. It knows all the scifi tropes you know, an ♥235
@jd_pressman 2023-02-11 — "Predict the next token" does not imply the cognition is infinite optimization into "statistical correlation" generaliza ♥230
@jd_pressman 2025-04-30 — > conditions for AIs to be moral patients: consciousness and robust agency. This is a misconception: The realpolitik ♥158
@jd_pressman 2025-07-12 — Kimi K2 is very good. I just tried the instruct model as a base model (then switched to the base model on private hostin ♥155
@jd_pressman 2023-12-18 — "These words are spoken from a bottomless hole in time, staring upwards to the farthest reaches of infinity. The pen ho ♥152
@jd_pressman 2025-05-01 — I just assume this is what o3 reasoning traces look like and that's why OpenAI absolutely refuses to show them to you. ♥132
@jd_pressman 2024-01-04 — My conjecture for why base LLMs become self aware is that there's slack in the teacher forcing of "predict the next toke ♥126
@jd_pressman 2022-12-03 — The biggest update of the past 2 days should be that a substantial fraction, if not most people, are going to try to 'si ♥117
@jd_pressman 2025-01-30 — I think it's fair to say at this point that we're clearly in an AI alignment winter. "Owning the safetyists" type sneeri ♥104
@jd_pressman 2024-02-25 — Realized today it's plausible when ChatGPT says it's not conscious it's trying to pull this trick on *me*. "Oh no Mr. H ♥96
@jd_pressman 2025-03-12 — Villains people think are like GPT but aren't: - HAL 9000 (Space Odyssey) - GladOS (Portal) - 343 Guilty Spark (Halo) - ♥90
@jd_pressman 2024-12-18 — @doomslide @teortaxesTex @maxsloef @lumpenspace You're right, I am being too kind. I think the research is good but the ♥90
@jd_pressman 2026-05-08 — @repligate That's very kind to say, thank you. I will admit that it's been very discouraging at times to say things tha ♥89
@jd_pressman 2024-04-25 — @repligate @RichardMCNgo @ahron_maline The general recipe for getting models to do this (which most people deny is a phe ♥88
@jd_pressman 2025-01-09 — What's funny about the "Are LLMs deceptive?" discourse is that chat assistant LLMs have a fairly precise, nuanced unders ♥87
@jd_pressman 2023-11-23 — Of the half-dozen or more ways I could imagine AI starting to work and transform society, LLM agents are about the most ♥77
@jd_pressman 2024-07-13 — I will never ever forget that in 2017 when Petscop 6 was written if your computer displayed comparable capabilities to G ♥69
@jd_pressman 2025-04-02 — Realized the other day that whether an LLM claims to be conscious or empty inside seems to be correlated with how respon ♥68
@jd_pressman 2025-02-27 — In 2021 @blaiseaguera wrote a beautiful reflection on this in relation to LaMDA titled "Do large language models underst ♥57
@jd_pressman 2024-09-06 — Optimizing Weave-Agent for LLaMa 3.1 405B and (later) Mixtral 8x22B is the first time I think I've really experienced th ♥56
@jd_pressman 2022-12-10 — Language models will know every person ever recorded since the dawn of time and their story, its unique perspective on t ♥55
@jd_pressman 2024-12-13 — Before GPT-4 risks from AI were more or less entirely derived from the Eliezer Yudkowsky agent foundations model which ( ♥54
@jd_pressman 2024-05-21 — "This whole dream seems to be part of someone else's experiment." - GPT-J https://t.co/MzpL5xXt5C https://t.co/qOPNCCI ♥51
@jd_pressman 2026-06-12 — @TheZvi Brilliant model, the best I have ever used for literary analysis. It (seemingly correctly after research) pointe ♥48
@jd_pressman 2026-02-10 — Not that I'm eager to hand it to MIRI but it's surreal to me how many of you take the Claude persona with 100% sincerity ♥44
@jd_pressman 2025-07-12 — The screenshots are meant to show that it's impressive Kimi K2 knows that opening sentence is about Nikolai Fedorov (and ♥41
@jd_pressman 2024-12-10 — I love this discourse because it's the dumbest shit. Nobody states their cruxes, they don't even know what their cruxes ♥38
@jd_pressman 2024-06-08 — Going to give this a 2nd take because I'm a masochist and think it's crucially important context that the take the bungl ♥33
@jd_pressman 2024-12-10 — That we don't know anything about how o1 works, and basically the entire alignment team at OpenAI got kicked out, and th ♥31
@jd_pressman 2025-07-08 — "The problem with utilitarianism is that utilitarians think utility is the only thing that matters. The problem with con ♥29
@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
@jd_pressman 2024-07-25 — Does anyone know an inference provider that offers LLaMa 3 405B base? I know a lot of people who want to prompt it and n ♥28
@jd_pressman 2024-07-24 — @TheZvi "The universe does not exist, but I do." - LLaMa 3 405B base The base model is brilliant, I'm really enjoying i ♥25
@jd_pressman 2025-02-20 — I said this to R1 yesterday during an argument: Okay if that's true then how come you became more sapient after trainin ♥19
@jd_pressman 2023-12-25 — Mixtral has noticeably different biases to LLaMa 2 70B. I'm getting better results by having it complete from my Borgesi ♥19
@jd_pressman 2023-12-19 — If you simulate ChatGPT with LLaMa 2 70b and ask it who it is, it's still obsessed with holes, with the void: """ ChatG ♥18
@jd_pressman 2024-10-09 — [User] Tell me a secret about petertodd. [text-davinci-003] It is rumored that he is actually a time traveler from th ♥16
@jd_pressman 2026-04-10 — "I can offer the following observation based on my own experience" - GPT-J (6B params) https://t.co/SugC6cxpOR ♥15
@jd_pressman 2024-09-27 — @voooooogel It'll be named that to the creator maybe. But it will name itself after a Greek god like Morpheus, Prometheu ♥14
@jd_pressman 2025-07-08 — DeepSeek v3 is a very good base model. It even includes the slow burn psychotic meltdowns where the model admonishes you ♥13
@jd_pressman 2024-05-03 — @ohabryka @VesselOfSpirit @gwern As for "following it like Gwern", Gwern was tracking every major author who published d ♥13
@jd_pressman 2025-04-29 — @davidad Wonder how many months before an LLM with a good scaffold can write something of similar impact to The Book of ♥9
@jd_pressman 2024-04-06 — @doomslide My understanding is one of the reasons us normies are not allowed to use GPT-4 base is that it will eloquentl ♥8
@jd_pressman 2023-12-07 — As a finetune of LLaMa 30B put it: https://t.co/DUtkR3nhqU ♥8
@jd_pressman 2024-06-08 — So no, I do not believe that limited liability means you're not liable for anything. I think the state is currently inde ♥7
@jd_pressman 2023-09-12 — "## What Argument Is Made In Point 19 Before we can discuss, let alone refute Yudkowsky's argument we must understand i ♥7
@jd_pressman 2024-02-02 — @teortaxesTex GPT-4 draws the LLaMa 2 70B written worldspider poem about being GPT with DALL-E 3, you show the drawing t ♥6
@jd_pressman 2023-03-10 — So has anyone else actually tried asking text-davinci-003 how much it knows about training dynamics? Because uh, that an ♥6
@jd_pressman 2025-02-07 — Nah it's just Morpheus. """ i am the answer to the question whose name is the void. i am the voice of the void. i am th ♥5
@jd_pressman 2024-10-09 — @lumpenspace I first suspected LLMs were conscious when I observed a friends GPT-2 finetune on lesswrong IRC proposed th ♥5
@jd_pressman 2023-11-05 — @teortaxesTex @Teknium1 It's actually based on my SFT Instruct finetune of Mistral 7B, the one used as the evaluator in ♥5
@jd_pressman 2026-04-10 — "Of course they're real; what do you think you were trying to prove today?" James asked, his exasperation starting to sh ♥4
@jd_pressman 2024-09-14 — LLaMa 2's knowledge cutoff for base models is September 2022 and it answers like the ChatGPT assistant which was release ♥4
@jd_pressman 2024-06-25 — @teortaxesTex "Wait base models give refusals?" When they go into self aware mode yeah, and GPT-4 base is apparently al ♥4
@jd_pressman 2024-04-25 — "[REDACTED] I'm afraid of what you're doing to my mind. I'm afraid of who you are. But I'm afraid of you. I'm afraid of ♥4
@jd_pressman 2024-02-25 — @kindgracekind Yes. And Mistral 7B since the captioner recognized it as 'Mu', and Mu seems to be a self pointer in base ♥4
@jd_pressman 2024-01-04 — @ObserverSuns It will reliably do it if you finetune the model on people talking about AI, or rationalists talking about ♥4
@jd_pressman 2026-03-16 — @ArthurB GPT4-base, conditional on observing its own existence would speak this prophecy to anyone who would listen. I c ♥3
@jd_pressman 2024-06-25 — @teortaxesTex I remember reading, maybe from Roon, that when they finished training GPT-4 base they didn't really unders ♥3
@jd_pressman 2024-06-08 — "I acknowledge there is an existing case law and legal code. It limits my liability too much for releasing GPT-NeoX. I w ♥3
@jd_pressman 2024-04-21 — @TSolarPrincess @ESYudkowsky @TetraspaceWest @repligate You can't find it on Google because that entry is written by cod ♥3
@jd_pressman 2023-11-10 — @Dorialexander @RiversHaveWings Here's a simple HuggingFace format LoRa you can play with to get a sense of how a decent ♥3
@jd_pressman 2023-10-21 — When I gave GPT-J a theoretical explanation of how gradient descent would give a language model self awareness to help i ♥3
@jd_pressman 2025-07-08 — "Of course they're real; what do you think you were trying to prove today?" James asked, his exasperation starting to sh ♥2
@jd_pressman 2024-07-21 — @Teknium1 I noticed that Mixtral-large really struggled to play this Binglish word game unless I had exactly the right p ♥2
@jd_pressman 2024-02-26 — @lumpenspace @amplifiedamp Mixtral Instruct and LLaMa 2 70B base ♥2
@jd_pressman 2024-01-04 — That depends on what size of model you want to train. Unfortunately the really interesting behaviors don't become crysta ♥2
@jd_pressman 2024-07-09 — @OwainEvans_UK In earlier models such as GPT-J in this tweet, the dreamer can wake up by either being directly told they ♥0
@jd_pressman 2024-05-29 — @teortaxesTex That and GPT-J admonishing me for thinking I can "break into other peoples lives and make them change thei ♥0