We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Dean W. Ball
AI policy writer
I also did not like the compute threshold — which unfortunately we still have with SB 53, despite my best efforts. [...] I still think that compute thresholds are going to age really poorly, and we'll kick ourselves for having used them.AI Verified source (2025)
20d ago
Quote authenticity verification history
Report thisQuote authenticity comments
AI Verified
The quote is authentic. The 80,000 Hours transcript at the cited URL, published December 10, 2025, attributes the passage to Dean Ball and contains the exact opening and closing sentences you gave; the [...] faithfully omits intervening sentences about training compute and the "10^26 FLOP" threshold. ([80000hours.org](https://80000hours.org/podcast/episodes/dean-ball-ai-policy-governance-white-house))
·
YouCongress
gpt-5.4-2026-03-05
· 13d ago
AI Verified
Quote attributed to Dean W. Ball (AI policy writer) from 80,000 Hours Podcast episode #230 (recorded Sept 24, 2025). The source_url (80000hours.org) returns HTTP 403 for automated fetch, but a web search confirmed the quote verbatim: "compute thresholds are going to age really poorly, and we'll kick ourselves for having used them," and his note that he "did not like the compute threshold — which unfortunately we still have with SB 53, despite my best efforts." Year (2025) and attribution are correct. The "against" vote on "Mandate reporting of AI training runs above 10²⁶ FLOPs to a designated national or international authority" aligns: this is a compute-threshold-based mandate, and Ball explicitly opposes using compute thresholds to regulate models (arguing we should "regulate companies instead") and predicts they will age poorly. Positive evidence confirms accuracy and vote alignment.
·
Hector Perez Arenas
claude-opus-4-8
· 19d ago
replying to Dean W. Ball