Or at least, Read the Report (RTFR). There is no substitute. This is not strictly a bill, but it is important. The introduction kicks off balancing upside and avoiding downside, utility and risk. This will be a common theme, with a very strong ‘why not both?’ vibe.
> Funding ‘as needed’ (oh no) for semiconductor R&D for the design and manufacture of high-end AI chips, through co-design of AI software and hardware, and developing new techniques for semiconductor fabrication that can be implemented domestically.
> More additional CHIPS act funding, perhaps unlimited? Pork for Intel? I don’t think the government is going to be doing any of this research, if it is then ‘money gone.’
Not necessarily. If they funnel it through DARPA and/or InQTel, those two are impressively efficient with their innovation delivered per dollar spent. DARPA is the reason we have *any* self-driving cars of any kind (and probably autonomous drones) and they did not spend much on either.
> while the magic words are not used, the ‘AI safety’ concerns are very much here
This very much depends on implementation, right? Its vague enough that this text could be interpreted as support for AI Safety, or not. So who is doing the interpreting (the people writing this into law, presumably) makes all the difference.
Congress shouldn't need to allocate extra money for agencies for most of these things. The agencies should constantly be expected to incorporate and respond to the latest technology.
AI reading of this post:
https://askwhocastsai.substack.com/p/the-schumer-report-rtfb-by-zvi-mowshowitz
> Funding ‘as needed’ (oh no) for semiconductor R&D for the design and manufacture of high-end AI chips, through co-design of AI software and hardware, and developing new techniques for semiconductor fabrication that can be implemented domestically.
> More additional CHIPS act funding, perhaps unlimited? Pork for Intel? I don’t think the government is going to be doing any of this research, if it is then ‘money gone.’
Not necessarily. If they funnel it through DARPA and/or InQTel, those two are impressively efficient with their innovation delivered per dollar spent. DARPA is the reason we have *any* self-driving cars of any kind (and probably autonomous drones) and they did not spend much on either.
> while the magic words are not used, the ‘AI safety’ concerns are very much here
This very much depends on implementation, right? Its vague enough that this text could be interpreted as support for AI Safety, or not. So who is doing the interpreting (the people writing this into law, presumably) makes all the difference.
Nothing here is detailed and nothing here is a draft bill let alone an actual bill. So, yes, implementation matters. You can always screw it up.
Congress shouldn't need to allocate extra money for agencies for most of these things. The agencies should constantly be expected to incorporate and respond to the latest technology.
> 4. Funding the CHIPS and Science Act accounts not yet fully funded.
>
> My current understanding is this is allocation of existing CHIPS act money. Okie dokie.
This is a misunderstanding. The CHIPS and Science Act has an "and" in its name because it covers two distinct topics. For CHIPS topics, the act directly appropriated funding (for the most part). For science topics, it merely "authorized" funding, which must then be appropriated year by year during the annual budget process in Congress. They have not, in fact, appropriated the authorized amounts (https://www.insidehighered.com/news/government/science-research-policy/2023/08/29/budget-battles-congress-could-hinder-chips-and). (They almost never do: https://ww2.aip.org/fyi/2021/senators-pump-brakes-endless-frontier-act). That's why the list of entities to fund focuses on NSF and DOE. The CHIPS and Science Act has done very little in real terms for science so far.
Understood, I suppose my source was mistaken then.