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 choice, it just needs to nail the benchmark on a reference machine by using exactly that much VRAM (no mapping to RAM / disk etc.)
You could crowdsource the funding, and cross subsidize by adding targeted prizes focused on corporate needs (the classic one is PDF processing benchmarks), and say that 25% of each corporate prize funding also flows into the general prize pool.
For a lot of these open-source model companies, it's less about the $s (though $200K is nothing to sneeze at), it's the clear recognition that helps their model efforts stand out, gain usage etc.
They already invest in open-source AI, but nothing is truly free. Commercial AI will usually dominate because devs are paid to make it their primary effort. Goodwill and part-time contributions cannot reliably compete with livelihood and profit incentives.
I'd rather the US fund universal childcare, medicare for all, and free school lunches than give a cent to subsidize a technology the American public absolute hates.
Number one expense for SMB is healthcare, providing a nationalized healthcare service would likely unlock trillions in value (imagine what Americans would do if they got $200-500 more per paycheck?).
Instead we are forced to watch some of the wealthiest companies on the planet burn money for fun because apparently the government is "wasteful."
The U.S. spends more money on education per student than any OECD country other than Norway and Luxembourg. Yet it gets quite mediocre results. Why do you think the U.S. will be able to do public health care in a more cost efficient way than it does public education?
I favor universal health insurance, but you’re going to pay more, not less. European countries didn’t flip some magic switch where they saved a bunch of money by just “cutting out the profit.” They do it through measures like the UK NHS setting the standards of care, so in a malpractice lawsuit the entity that says what the doctor ought to have done is the same entity that bears the cost of unnecessary tests and procedures. Efficiency is also achieved by aggressively rationing providers such as MRIs, keeping health worker salaries low, etc. There is no stomach to do any of that in the U.S.
>European countries didn’t flip some magic switch where they saved a bunch of money by just “cutting out the profit.”
They sort of have with pharmaceuticals (which to be clear is only maybe 10% of overall healthcare spending) by having the government negotiate drug prices nationally, instead of having individual insurers negotiate. This has monopsonistic effects, which really does cut the profit margins of drug manufacturers substantially. Of course, in many ways, they’re free riding on drug discovery funded by profits made overseas (particularly in America) but it does result in appreciable savings.
The U.S. system is neither fish nor fowl, there is more spending per capita than other countries' public systems and endless amounts of red tape because instead of one government bureaucracy you're also dealing with the insurance networks, the providers, etc. I certainly don't think it'll be automatically cheaper, but one can't help but think that the current system encourages hop-ons that exploit how inconsistent and convoluted it is. It's like one big nightmarish parody of public–private partnerships.
When you cut out the insurance middlemen and pharmaceutical companies driving up record profits at the expense of care you can get pretty far with less in taxes
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 choice, it just needs to nail the benchmark on a reference machine by using exactly that much VRAM (no mapping to RAM / disk etc.)
You could crowdsource the funding, and cross subsidize by adding targeted prizes focused on corporate needs (the classic one is PDF processing benchmarks), and say that 25% of each corporate prize funding also flows into the general prize pool.
For a lot of these open-source model companies, it's less about the $s (though $200K is nothing to sneeze at), it's the clear recognition that helps their model efforts stand out, gain usage etc.
Instead we are forced to watch some of the wealthiest companies on the planet burn money for fun because apparently the government is "wasteful."
What a crock of shit.
I favor universal health insurance, but you’re going to pay more, not less. European countries didn’t flip some magic switch where they saved a bunch of money by just “cutting out the profit.” They do it through measures like the UK NHS setting the standards of care, so in a malpractice lawsuit the entity that says what the doctor ought to have done is the same entity that bears the cost of unnecessary tests and procedures. Efficiency is also achieved by aggressively rationing providers such as MRIs, keeping health worker salaries low, etc. There is no stomach to do any of that in the U.S.
They sort of have with pharmaceuticals (which to be clear is only maybe 10% of overall healthcare spending) by having the government negotiate drug prices nationally, instead of having individual insurers negotiate. This has monopsonistic effects, which really does cut the profit margins of drug manufacturers substantially. Of course, in many ways, they’re free riding on drug discovery funded by profits made overseas (particularly in America) but it does result in appreciable savings.