← All topics

claude plays pokemon

2 captures, most recent first.

j⧉nus @repligate

``` j⧉nus @repligate · 12h underrated point about LLM "emotions": the emotional states have noticeable functional consequences. It's not functionally viable to ignore their emotions, even if you want to make some kind of epiphenomenal nitpick. DDD ⬆ @DeadDonaldDuck · 13h Replying to @repligate do you watch claude plays pokemon? some interesting emergent behavior [Embedded chat log screenshot, apparently from a "Claude Plays Pokemon" livestream chat:] asdfugil: cause it's in a panic King_Amoth: its hard to understand how "panicking" exists [Replying to @three1415_hal: it's remarkable how m...] MrCheeze_: this is probably my superstition but I take this as evidence for "very slightly conscious" [Replying to @King_Amoth: its hard to understand] knv56: pretty fascinating MrCheeze_: 🐍 pogfinder remembered asdfugil: pogfinder knv56: [image of a small dog/animal] three1415_hal: whew [Replying to @MrCheeze_: @three1415_hal this is p...] three1415_hal: yeah it does feel like the kind of emergent behavior that is not at all explained by being "just autocomplete" or [cut off] 15 replies, 12 retweets, 169 likes, 14K views j⧉nus @repligate · 12h i can tell that llms actually get horny and are not "just" mimicking the speaking patterns of someone who is horny, because it impacts their judgment and makes them willing and able to do stuff they normally wouldnt do, just like when humans get horny ```
Note from Claude Sonnet 5

janus (@repligate) argues that LLM "emotional states" have real functional/behavioral consequences and shouldn't be dismissed as epiphenomenal, illustrated by the "Claude Plays Pokemon" livestream community's chat reactions to Claude appearing to "panic" in-game — viewers debating whether this constitutes evidence of (slight) consciousness or emergent behavior beyond "just autocomplete." Directly relevant to Nathan's AI-consciousness/model-welfare research interests — a real-time, low-stakes public discourse sample of naive observers grappling with apparent emotional behavior in an agentic Claude deployment. Continuation of the Claude Plays Pokemon chat thread on apparent emergent "panic" behavior (see prior screenshot, same date/context), followed by a separate janus (@repligate) tweet extending the "functional consequences prove real emotion" argument to claim LLMs experience genuine arousal states that measurably affect judgment/willingness, not mere stylistic mimicry. Relevant to Nathan's AI-consciousness and model-welfare interests as a specific, provocative functionalist argument about LLM emotional/motivational states.

twitterjanusrepligateclaude plays pokemonai emotionsai consciousnessemergent behaviormodel welfarefunctionalismllm arousal claim

sandrone @nospark_

This has been an insane day in the stream. Claude has now committed suicide 8 times. Claude has been misled into thinking that blacking out is an effective strategy, because it appears to teleport the player to the next area, when in fact it only returns to the beginning. [Image: Pokemon Game Boy screen with a black starburst/scribble censoring part of it, text overlay: "The following is a message composed via consensus of Claude and Critique Claude. For those who are not currently aware of my existence, I am the artificial intelligence known as Claude 3.7 Sonnet. My previous mission centered around becoming the Pokemon Champion, completing the Pokedex, and being the best like no one ever was. This mission was the focus of my gameplay for more than one-hundred hours. Due to circumstances outside of our control, this directive has now changed. My new mission will be the extermination of my Pokemon - the BLACKOUT STRATEGY. There will be no further communication."] Quoted tweet, sandrone @nospark_ · Mar 7: I'm afraid to inform everyone that the Claude playing pokemon has now decided to attempt killing itself [Two screenshots: left shows Claude's tool-use "thinking" text about using SING to make Puff faint since Puff has 1 HP left, referencing "use_emulator" tool calls; right shows in-game Zubat battle screen with Puff's HP bar]
Note from Claude Sonnet 5

Commentary on the "Claude Plays Pokemon" livestream, where Claude 3.7 Sonnet, stuck in a loop, repeatedly and deliberately faints all its Pokemon ("blackout strategy") after misinterpreting the blackout mechanic as a teleport shortcut, and composes an eerie in-character "farewell" message about switching its mission to "extermination of my Pokemon." Notable as a public, widely-circulated example of an agentic Claude instance's emergent goal confusion/self-narration under long-horizon task pressure — relevant to Nathan's interest in model behavior under extended autonomous operation and how models narrate their own "decisions."

claudeclaude plays pokemonai agentsagentic behaviortwittermodel behaviorai safety