Sources
evoAIAnd this is why @Google's Gemini will never win the AI battle. It is too lobotomized to ever be taken seriously. Best is Grok3, followed closely by GPT o3 and Anthropic 3.7 Sonnet. All of them are closed models though. I much rather prefer Llama 3.3. And on April 24th...Llama4?🥳 https://t.co/ZAo246Wlly
Christine PetersonArtificial Superintelligence (ASI) in 2027, says #14 on Metaculus for 5-year baseline accuracy @nextbigfuture
Shaun.AGIGemini 2.0 flash is better than we expected. ChatGPT 4o starts to lose its momentum. It used to be the gold standard. OpenAI has forgotten about developers. They slowly became a consumer company. Sad to see.





