Sign in

iiinova is part of the Atmosphere — an open, public network where you own your identity.

Dallas

@dallasthe.dev
133 followers 268 following 25 posts

SRE/DevOps London He/Him

PostsRepliesMedia
Dallas @dallasthe.dev · 02/03/2025
About 12 hours later and Claude managed to get Mistys gym badge at some point, now it’s just stuck in Cerulean city going round and round. I wonder if in a year we’ll have 10m token context windows and this sort of problem will seem like a hillarious Stone Age issue to us 🤔
ClaudePlaysPokemon aimlessly wondering around Cerulean City
110 1
Dallas @dallasthe.dev · 01/03/2025
I'ts just going in and out of the pokemon center now, I think I'm going to bed...
100 1
Dallas @dallasthe.dev · 01/03/2025
Based on the graph they released the twitch stream is maybe a little behind, if it can get Mistys badge first time it'll be back on track (If I'm remembering the timeline for pokemon correctly)
Claude 3.7 Sonnet demonstrates that it is the very best of all the Sonnet models so far at playing Pokémon Red. On the x-axis is the number of interactions Claude completes as it plays the game; on the y-axis are important milestones in the game involving collecting certain items, navigating to certain areas, and defeating certain game bosses.
110 1
Dallas @dallasthe.dev · 01/03/2025
It mmadddee itttt
Claude makes it to the pokemon center
110 1
Dallas @dallasthe.dev · 01/03/2025
Nevermind we're back in the bike shop
Screenshot of twitch.tv/ClaudePlaysPokemon, the LLM has gone back to the bike shop it chose to leave earlier and has realised, stating "I've made a crucial mistake"
100 1
Dallas @dallasthe.dev · 01/03/2025
It's pretty fixated on the pokemon center, managed to get out of the bike shop quickly and avoided getting distracted, arguably better than me in that regard
100 1
Dallas @dallasthe.dev · 01/03/2025
Claude's been "choosing" to nickname all the pokemon, I think that's interesting maybe ?
Claudes pokemon team, the LLM has nicknamed all the pokemon instead of use the default name
110 1
Dallas @dallasthe.dev · 01/03/2025
Claude plays Pokemon is taking off right now, after 78 hours it’s made it out of mt moon
Screenshot of twitch.tv/claudeplayspokemon, the LLM has been in mt moon for 78 hours A screenshot of twitch.tv/ClaudePlaysPokemon, the LLM has made it outside mt moon
161 6
Dallas @dallasthe.dev · 13/01/2025
It’s not localised but it seems to have ‘localised’ cards on the bbc app home page weirdly, very odd
230 3