Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Governments, companies, nonprofits should invest in free, open source AI [pdf]

siegelendowment.org

Post reply on HN
Print view View original

61–70 of 134 posts

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#61

Post by NitpickLawyer »

rao-v wrote: post #8

We really need to band together to fund / sponsor targeted inducement prizes (a la Nobel laureate Michael Kremer) for open models. Every 6-12 months, give out $200K to the first model to hit a min threshold on a set of ~5-10 hard benchmarks (+ perhaps one secret benchmark) using a total of 16GB / 32GB / 64GB / 128GB of VRAM (at a min context length of 200K), then move the threshold up. Quantization etc. is dealers ch…

I think the Korean government did have a competition like this, I remember last year we got a bunch of models released at the same time to make the cut for the next stage. The models weren't anything to write home about, IMO.

Having it with clear hw requirements tiers is a nice differentiator. The only issue is that the benchmarks would 100% need to be closed, no other way around it. And then you have the issue of creating and curating good evals for every "stage" of the project. That's a hard task even for "honest" lab-internal evals. And you'd have to publish those evals after each round (for trust purposes), and start over for the next round. Doable, but it would cost a lot (probably more than the prize pools) and you'd have to keep doing this.

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#62

Post by NitpickLawyer »

thoughtpeddler wrote: post #59

Ah, this from the same David Siegel who said almost 2 yrs ago (in a talk found here: https://youtu.be/0z60xUDo-NI?si=PTDe11-sn2P53qo5&t=420 ) that the AI data center buildout was premature because: > Even if the current approaches will continue to scale, this would be as if in the early days of computing, perhaps someone invented a bubble sort for sorting numbers (an n-squared algorithm), and the tech companies at th…

> AI data center buildout was premature

Ask Amodei how he feels about going to spaceman bad for compute that he couldn't find anywhere else in the market.

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#63

Post by simianwords »

boosturpud wrote: post #21
simianwords wrote: post #3

Earlier quoted context omitted.

Redistribution can only get you so far. Creating new wealth is more sustainable.

> Redistribution can only get you so far. That "so far" being a middle class the envy of the rest of the world which the US threw away to create a new class of oligarch. "Back at the tail-end of that era, in the early 1960s, America’s richest faced a 91 percent tax rate on income in the top tax bracket." https://inequality.org/article/tax-the-rich-we-did-that-once...

Perhaps they realised it’s not a good policy? Like literally everywhere else in the world?

No other country has 91% taxes and all those who tried dialled it back. India tried it and it was hilariously bad.

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#66

Post by BrtByte »

rao-v wrote: post #8

We really need to band together to fund / sponsor targeted inducement prizes (a la Nobel laureate Michael Kremer) for open models. Every 6-12 months, give out $200K to the first model to hit a min threshold on a set of ~5-10 hard benchmarks (+ perhaps one secret benchmark) using a total of 16GB / 32GB / 64GB / 128GB of VRAM (at a min context length of 200K), then move the threshold up. Quantization etc. is dealers ch…

I would just add a reproducibility requirement and avoid keeping the exact same benchmarks for too long

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#67

Post by BrtByte »

NitpickLawyer wrote: post #61
rao-v wrote: post #8

We really need to band together to fund / sponsor targeted inducement prizes (a la Nobel laureate Michael Kremer) for open models. Every 6-12 months, give out $200K to the first model to hit a min threshold on a set of ~5-10 hard benchmarks (+ perhaps one secret benchmark) using a total of 16GB / 32GB / 64GB / 128GB of VRAM (at a min context length of 200K), then move the threshold up. Quantization etc. is dealers ch…

I think the Korean government did have a competition like this, I remember last year we got a bunch of models released at the same time to make the cut for the next stage. The models weren't anything to write home about, IMO. Having it with clear hw requirements tiers is a nice differentiator. The only issue is that the benchmarks would 100% need to be closed, no other way around it. And then you have the issue of cr…

Yeah, I suspect the $200K check would be the cheapest line item in the whole thing

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#68

Post by steve_j_choi »

NitpickLawyer wrote: post #61
rao-v wrote: post #8

We really need to band together to fund / sponsor targeted inducement prizes (a la Nobel laureate Michael Kremer) for open models. Every 6-12 months, give out $200K to the first model to hit a min threshold on a set of ~5-10 hard benchmarks (+ perhaps one secret benchmark) using a total of 16GB / 32GB / 64GB / 128GB of VRAM (at a min context length of 200K), then move the threshold up. Quantization etc. is dealers ch…

I think the Korean government did have a competition like this, I remember last year we got a bunch of models released at the same time to make the cut for the next stage. The models weren't anything to write home about, IMO. Having it with clear hw requirements tiers is a nice differentiator. The only issue is that the benchmarks would 100% need to be closed, no other way around it. And then you have the issue of cr…

It was a competition organized by the Korean government but the directive wasn't for the same cause as the writer. It was more for constructing Sovereign AI for the country. Also, all models except Exaone had some weights copied from Chinese models, and from the corporate point of view, developing from-scratch model is not cheap despite the financial support from the government.

Yes, I hope the open model communities will someday be able to run current frontier models which will be able to handle autonomous tasks and the hardware to run it will be served at consumer level; however, like how recycling isn't profitable, no companies will fully commit to the movement. Don't get it twisted, I don't have a solution but maybe a global scale movement to liberate knowledge-library could suffice.

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#69

Post by SideburnsOfDoom »

simianwords wrote: post #63
boosturpud wrote: post #21

Earlier quoted context omitted.

> Redistribution can only get you so far. That "so far" being a middle class the envy of the rest of the world which the US threw away to create a new class of oligarch. "Back at the tail-end of that era, in the early 1960s, America’s richest faced a 91 percent tax rate on income in the top tax bracket." https://inequality.org/article/tax-the-rich-we-did-that-once...

Perhaps they realised it’s not a good policy? Like literally everywhere else in the world? No other country has 91% taxes and all those who tried dialled it back. India tried it and it was hilariously bad.

> Perhaps they realised it’s not a good policy?

Citation needed that it's not a good policy. Parent comment explained why it was good, and your reply is "or perhaps it was actually bad".

There's a simpler explanation as to why it was rolled back: the people with the most money and power didn't like it and wanted it gone, and worked over time to accomplish this.

Re: Governments, companies, nonprofits should invest in free, open source AI [pdf]

#70

Post by simianwords »

SideburnsOfDoom wrote: post #69
simianwords wrote: post #63

Earlier quoted context omitted.

Perhaps they realised it’s not a good policy? Like literally everywhere else in the world? No other country has 91% taxes and all those who tried dialled it back. India tried it and it was hilariously bad.

> Perhaps they realised it’s not a good policy? Citation needed that it's not a good policy. Parent comment explained why it was good, and your reply is "or perhaps it was actually bad". There's a simpler explanation as to why it was rolled back: the people with the most money and power didn't like it and wanted it gone, and worked over time to accomplish this.

so it could be two reasons

1. 90% tax is a net good for society but big bad capitalists colluded to reduce it in every country independently

or

2. 90% is actually a bad policy and every country independently adopted against it including Nordic countries, China, India, USA

you decide which one is more likely.

Post reply on HN