Will the GenAI scaling craze see a group spend >$1B to train a single large-scale model in the next 12 months?
Basic
10
Ṁ207Oct 2
52%
chance
1D
1W
1M
ALL
Taken from the State of AI Report 2023.
The GenAI scaling craze sees a group spend >$1B to train a single large-scale model.
This question will be resolved based on the resolution of the 2024 report.
Get
1,000
and1.00
Sort by:
@RemNi The market is about price, not parameter count. Unless you’re claiming that we know that llama 3 cost >$1B?
1 h100 = 40k dollars
350 000 x 40 000 = 14 000 000 000
Assuming training llama 3 is 1 / 20th of the compute they will use that cluster for, and then considering all other costs beyond just the GPUs, yeah easy 1B dollars
@RemNi Thanks for the info, though I don’t think that’s convincing enough for resolution. I think it might make sense to wait for an official number or an estimate by Epoch AI before resolving.
Related questions
Related questions
In the next 12 months will the GenAI scaling craze see a group spend >$1B to train a single large-scale model?
22% chance
SoAIR '23 5/10: Will the GenAI scaling craze see a group spend >$1B to train a single large-scale model by Oct 24?
17% chance
$1T AI training cluster before 2031?
63% chance
$100B AI training cluster before 2029?
68% chance
AI: Will someone train a $1B model by 2025?
66% chance
Will AI have a trillion+ dollar impact by the end of 2025?
29% chance
Will a large AI company (e.g. OpenAI) acquire an inference-focused AI chip compnay in the next 12 months?
62% chance
100GW AI training run before 2031?
53% chance
How many billion dollar AI training runs will occur in 2024?
AI: Will someone open-source a $100M model by 2025?
29% chance