I would imagine Anthropic is hitting a wall on the number of simultaneous agents that can be useful on a task and are working on increasing that maximum. I would assume if they could get twice as much progress by using 10000 simultaneous running fable instances instead their brief would state that.

IMO, it is very interesting and also very alarming. At some point the barrier to the exponential progress is the amount of computation available. I would consider hitting this limit and then likely using the same said swarm to help resolve it via various means as very much fitting under the definition of Ray Kurzweil's singularity.

Also, I would disagree, Fable is a sufficient coder. The average human being better, even if that is true now (and personally I wouldn't agree), the average human is limited in the number of 'human instances' he or she can interface with at one time.

if 1000 fable instances is better than any human at a task, how about 1 million instances running? Much easier than finding and hiring and coordinating 1000 extremely productive and intelligent human beings. Much cheaper too, it appears.

▲

a_wild_dandan 6 hours ago | parent [-]

10k OpenAI agents solved a Millennium Prize problem. Anthropic ain't gonna hit a wall there.

▲

bryzio 4 hours ago | parent [-]

Could someone explain how increasing number of agents help? Is each agent researching a potential solution space where there's many potential solution spaces? I'm struggling to conceptualize how a large agent swarm can be used to solve a research problem.

▲

pvab3 3 hours ago | parent | next [-]

I don't think it's nearly as relevant as the total amount of compute being used. I just think they like talking about swarms because it sort of anthropomorphizes them.

▲

genghisjahn 3 hours ago | parent | prev | next [-]

Feel like this might still apply to multi-agent programming:

I have earlier argued that the sheer number of minds to be coordinated affects the cost of the effort, for a major part of the cost is communication and correcting the ill effects of miscommunication (system debugging). The Mythical Man-Month Frederick P. Brooks Jr.

▲

bertmuir 3 hours ago | parent | prev | next [-]

By collaborating and division of labour.

	▲	Sharlin 2 hours ago \| parent [-]
		Similarly to how nine women can make a baby in one month by collaborating and division of labour.

▲

wyre an hour ago | parent | prev | next [-]

I am just as clueless as you are, but I think it having each agent work on the problem mostly individually in parallel, likely sharing their breakthroughs for the next wave of agents. I'm reminded of infinite monkey's writing Shakespeare; suggesting if you have an infinite amount of agents, they will end up solving a Millenium Prize Problem.

Here is OpenAI's blog post about Navier-Stokes; [1] which suggests they used smaller connected swarms given the different prompts, then the findings were used to write further follow up prompts for a new group of agents.

[1] https://openai.com/index/navier-stokes-solution/

▲

Tadpole9181 an hour ago | parent | prev [-]

Actually, I think Microsoft just posted the results of research the other day that would be relevant. They were looking at self-organized AI swarms instead of centralized orchestration, and found that just starting up thousands of SOTA models really can precipitate a natural hierarchy and work toward problem solving.

At the very least, the swarm would be able to set up a means of claiming avenues of thought, then verifying the solutions (or lack thereof) of other agents.

It's quite clear this process - whatever it may be - works, given the results we've seen in the past few weeks.