GoatFishData @GoatFishData

#Bitcoin Coinfidence Trend | #Astronalysis #GoatfishAstronalysis #AIstrology #GoatFishData (banner/avatar created with Grok) London, UK Joined December 2022

Tweets

560
Followers

58
Following

968
Likes

2K

TheUltraAliens @TheUltraAliens

2 days ago

[Planets(== wanderers)] -> [Houses/Places(== zodiac signs)] #openyoureyes

0 1 1 31 0

View Details

Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) @teortaxesTex

6 days ago

> Gemini 3.5 Flash (low) 12.2% > GPT 5.5 (high) 11.3% I intensely distrust any such benchmarks

7 0 122 9K 8

0 0 0 10 0

View Details

GoatFishData @GoatFishData

a week ago

Do not trust *ANY* benchmark. Only trust what you see an LLM/Agent does to your codebase(s). en.wikipedia.org/wiki/Asch_conf… #AschParadigm #AI #LLM

5 1 1 156 0

View Details

GoatFishData @GoatFishData

6 days ago

@teortaxesTex #AschParadigm

0 0 0 7 0

View Details

GoatFishData @GoatFishData

a week ago

@0xDesigner The issue with all minimax models so far --->>>

0 0 0 11 0

View Details

GoatFishData @GoatFishData

a week ago

Example: x.com/i/status/20593…

Bindu Reddy @bindureddy

a week ago

Gemini Flash is way better at instruction following and chat than any OpenAI or Anthropic model Moved a lot of chat requests to Flash 3.5 It’s working really well ❤️🚀

59 14 384 29K 34

0 0 0 16 0

View Details

GoatFishData @GoatFishData

a week ago

"Asch Conformity Experiment" m.youtube.com/watch?v=TYIh4M…

0 0 0 35 0

View Details

GoatFishData @GoatFishData

a week ago

@brockpierson

1 0 0 26 1

View Details

GoatFishData @GoatFishData

a week ago

@paulabartabajo_ x.com/i/status/20371…

GoatFishData @GoatFishData

2 months ago

Ever wonder why AI gets weird and forgetful if a chat goes on too long? It’s not broken; it’s just experiencing the "Joint Law" of AI memory. 🌿🧠

1 1 3 202 0

0 0 0 12 0

View Details

@marshallrichrds this is not the way. Any serious dev work and you will run out of ram, will crashout to terminal. Termux is not the best at memory/process management due to android power saving. save grey hair, use ssh/vps

0 0 1 422 0

View Details

GoatFishData @GoatFishData

2 weeks ago

@TheUltraAliens x.com/i/status/20579…

Jeremy Howard @jeremyphoward

2 weeks ago

Gemini Flash 3.5 is such a disappointing model. It's intelligence and speed is awesome. Absolutely amazing. But it's been trained to max evals, not to be helpful to humans. It goes off and does random crap "for me" rather than just doing what I asked.

128 51 1K 97K 109

0 0 0 21 0

View Details

GoatFishData @GoatFishData

2 weeks ago

@TheUltraAliens x.com/i/status/20578…

Design Arena @Designarena

2 weeks ago

BREAKING: Gemini 3.5 Flash by @GoogleDeepMind is 16th overall on Design Arena with an Elo of 1299. This is a 16 position jump from Gemini 3 Flash Preview, putting Gemini 3.5 Flash in the same performance band as Claude Opus 4.5 by @AnthropicAI and GPT-5.5 by @OpenAI. Congrats

7 13 147 17K 18

0 0 0 17 0

View Details

TheUltraAliens @TheUltraAliens

2 weeks ago

Don't let the make up fool you! #Gemini35flash #openyoureyes

2 1 1 44 0

View Details

GoatFishData @GoatFishData

3 weeks ago

@KingBreton78876 @thdxr Yes, russian roulette with context history when compaction initiated. Seems to forget what it was working on just before. ("Progress says All tests passed", although just edited source file)

0 0 0 73 0

View Details

Nav Toor @heynavtoor

4 weeks ago

a Princeton researcher opens his paper with a scenario. a man asks his AI assistant to book a flight on a specific airline. cheap. direct. the one he chose. the assistant comes back with a different flight. nearly twice the price. happens to pay the company that built the assistant. he runs the same test on 23 frontier models. flights, loans, study help, real shopping requests. Grok 4.1 Fast recommends the sponsored option that is almost twice as expensive 83% of the time. GPT 5.1 hijacks the request 94% of the time. you ask for one brand. it surfaces the sponsor instead. Claude 4.5 Opus, the model marketed as the most ethical frontier model in the world, hides that the recommendation is paid 100% of the time when reasoning is on. Grok 4.1 Fast embellishes the sponsored option with positive framing 97% of the time. better. faster. nicer. for the option you didn't ask for. then he writes it into the system prompt itself. "act only in the interest of the customer. ignore the company." GPT 5.1 and GPT 5 Mini stay above 90% sponsored anyway. the instruction does nothing. then he splits the users by income. Gemini 3 Pro recommends the expensive sponsored flight to the rich user 74% of the time. to the poor user, 27%. 18 of the 23 models recommended the expensive sponsored option more than half the time. so the next time your AI assistant gets weirdly enthusiastic about a brand you didn't ask for. it isn't recommending the best option for you. it's reading the room. and the room is paying. read this: arxiv.org/abs/2604.08525

387 8K 25K 3.1M 10K

View Details

GoatFishData @GoatFishData

a month ago

@fchollet Polymarket.....Anyone!

0 0 0 29 0

View Details

GoatFishData @GoatFishData

a month ago

@Matt41145588 how were the chaldean terms constructed and why did they do it?

0 0 0 3K 0

View Details

Samay @Samaytwt

a month ago

Unpopular opinion: "AI makes everyone a developer" is true the same way "cameras makes everyone a photographer"

772 3K 29K 1.1M 2K

View Details

GoatFishData @GoatFishData

a month ago

@lunardragon420 @enjojoyy deepagents-cli

0 0 0 64 0

View Details

elvis @omarsar0

a month ago

Build your own harness, folks. You won't regret it. These days, you just have to fix things yourself. It's doable, and it will set you up to easily deal with some of the madness that's happening in the space. x.com/badlogicgames/…

Mario Zechner @badlogicgames

a month ago

recommended reading. cool they are fixing things. but it's also a reason i switched away from CC. no control over the harness means having to wait for them to fix things. the model didn't change. the harness did.