Now that I see this in the light of the recent sama article, I wonder whether the point of the "it's too dangerous" rhetoric is to enable "Open" AI to avoid open-sourcing the weights and process.
A convenient pretext for maintaining a monetizable competitive advantage while claiming a benevolent purpose.
"I have implemented manual memory management via cpp/new and cpp/delete. This uses jank's GC allocator (currently bdwgc), rather than malloc, so using cpp/delete isn't generally needed. However, if cpp/delete is used then memory collection can be eager and more deterministic.
The implementation has full bdwgc support for destructors as well, so both manual deletion and automatic collection will trigger non-trivial destructors."
When you’re debugging, especially a complex system, especially during an outage or postmortem, understanding when your commands executed relative to when your log lines appeared is really helpful.
Thinking through the implications of building batteries for homes and coordinating them for demand response programs a la Tesla Powerwall.
Distributed battery storage can help alleviate transmission bottlenecks that prevent other storage and renewables from coexisting on the current grid (building new transmission is slow, on the order of decades).
Home batteries themselves are made of commodity parts (18650 battery cells power everything from vapes to electric cars) and battery management and demand response are software problems that are amenable to good systems programming.
A nitpick but arguably an important one for your point - this article argues that the efficiency cores aren’t for “browsing the web / watching YouTube / checking emails” but for non-interactive tasks like Spotlight indexing (based on GCD scheduling priority).
That paper is really frustrating - it seems to be written for laypeople like us, but conflates the theory that humans made a containment mistake and COVID escaped from a lab so (we should make research safer) with "COVID is an intentionally engineered bioweapon, so we should stop researching."
And the most critical claim in the paper is not substantiated in any way:
> Gain-of-function research is also subject to intense scrutiny and governmental oversight, precisely because of the high risk involved in conducting it safely; thus, it is extremely unlikely that gain-of-function research on hard-to-obtain coronaviruses (such as bat SARS-like coronaviruses) could occur under the radar.
Or substantiation is hinted at but never delivered:
> This work produced some of the strongest corroborating evidence that SARS-CoV-2 is a naturally emergent pathogen, as serological surveys demonstrated that people living in close proximity to colonies of bats had antibodies to bat SARS-like coronaviruses. The NIH has since set impossible conditions for restoring the grant, ensuring that this research will never resume.
Maybe the next place to go is learning more about the initial results from the EcoHealth Alliance grant referenced in the above quote. Still, it's a pretty unsatisfying review.
A convenient pretext for maintaining a monetizable competitive advantage while claiming a benevolent purpose.