Which of the following will be true about the next major AI breakthrough? (after test-time compute)
2
Ṁ1kṀ522030
57%
it will make AI models better at agentic tasks (such as OSWorld and SWE-bench)
54%
It will allow AI models to become much better at working with long-context
54%
It will allow AI models to learn at test time
50%
It will allow AI models to hallucinate much less
50%
It will allow AI models to use long context more efficiently
50%
It will make make AI models better at visual reasoning
50%
It will make AI models much better at answering competitive math questions
46%
It will make AI models better at making art (creative writing, visual art)
Whether something counts as a breakthrough is according to my subjective judgement. Some examples of things I consider breakthroughs:
Transformer architecture
RLHF
Test-time compute
This question is managed and resolved by Manifold.
Market context
Get 
1,000 to start trading!
1,000People are also trading
Related questions
Which of the following AI-related statements will happen first?
Which kind of computers will be the standard for AI models in 2034?
When will a US government AI run overtake private AI compute by FLOP?
Before 2030, will an AI complete the Turing Test in the Kurzweil/Kapor Longbet?
55% chance
Will AI pass Video Turing Test by 2030?
68% chance
What will be the top-3 AI tools in 2040?
In what year will there be an AI capable of passing a high-quality Turing test?
Will AI pass the Rube Goldberg Turing test by the end of 2028?
38% chance
Will AI figure out who has the highest processing speed to working memory ratio by 2030?
48% chance
Which of the following breakthroughs will Deepmind achieve by 2030?