I don't understand. You will ask it for something you want, and instead of needing to pay a the team of people with expertise, it will just do it for you.
Someone will still need to ask for something (or maybe it will just run autonomously making things it thinks we'll like) but it won't require any expertise in the field.
I can see two reasons American companies might want to train models they give away for free:
1) They sell compute: chips (Nvidia), data centers (AWS, Microsoft, Google, SpaceX, etc), or even end-user device manufacturers like Apple (e.x. M7 rumored to have 1.5TB of unified memory). If Jevon's paradox holds, then cheaper (or free) models means more demand. But compute is likely supply-constrained for years anyway.
2) Their product isn't AI but depends on AI being cheap, or they don't want competitors to capture that value, i.e. "commoditize your complement" https://gwern.net/complement
It probably doesn't make sense for these companies to invest a lot of money training models that will be obsolete in a few months anyway. When progress starts to plateau I'd expect more companies to start training models they give away for free.
> Crucially it's the responsibility of those who are making the claim here to bring evidence and definitions
Ok, I claim that if something draws reasonable conclusions to questions it hasn't previously seen by performing steps that look like reasoning, then it is reasoning.
You wouldn’t even take social security benefits despite (I assume) having been forced to pay social security taxes?
Even Ayn Rand took social security benefits. Some people point to that as hypocrisy, but I feel if the government forced you to pay into a program it is moral to take back from it (though not more than you contributed), even if you morally object to the program overall.
You’d probably feel differently if you were one of the 10 million (and growing) Starlink subscribers who previously had worse (or no) high speed internet access.
> perhaps will miss the most important discoveries of the universe because of such programs
Unlikely for a number of reasons, and actually a large commercial constellation could help with scientific discovery.
Satellite trajectories are predictable from published ephemerides. It’s trivial for telescope operators to exclude the small fraction of pixels in a frame occupied by a satellite. They already do this.
Furthermore, the same technology that makes it economical to put millions of satellites into LEO (e.g. Starship) also makes it cheaper to build larger telescopes in space where they’re unaffected by the earth’s atmosphere (and satellites in LEO).
If SpaceX spent a ton of resources developing Moon and Mars base plane they’d be criticized for not focusing on the launch vehicle. The railroad company also doesn’t necessarily need to be the ones building the towns.
Starship has experienced setbacks but isn’t a “massive failure” by any measure.
Again, “What I find silly is the certainty that critics have that SpaceX will fail”.
I doubt you’ve done enough investigation to conclude that. Watching a couple a YouTube videos about how data centers in space are dumb doesn’t count.
For one, it’s not one big hundred megawatt-scale datacenter, but many smaller rack-sized satellites.
Two, there are many trade offs SpaceX can make to reduce the overall cost. The radiated power is proportional to the 4th power of the temperature, so anything they can do to increase the temperature drastically reduces the size of the radiators, including custom silicon, heat pumps, etc. There’s also novel radiator technology like liquid droplet radiators that could be worth developing.
Third, there are reasons to put it in space: solar panel efficiency, no cost for land, less regulation, no NIMBYs, etc.
I’m confident you could devise a way to ensure the verifying party doesn’t know which website requested the verification, and vice versa. I’ll have to think about it a bit more.
In general I'm opposed to this kind of regulation, but as a thought exercise, we do have the primitives needed to do age (or any other attribute) verification in a privacy-preserving and decentralized way.
You could imagine a hierarchy of organizations (governments, financial institutions, schools, etc) that a website trusts to verify some attribute (minimum age, citizenship, etc). Those organizations can attest that some identifier like an email address has been verified to belong to a real individual with that attribute, and that organization belongs to the hierarchy the website trusts, without revealing anything else about the user, the exact verifying organization, or the requesting website.
Someone will still need to ask for something (or maybe it will just run autonomously making things it thinks we'll like) but it won't require any expertise in the field.