@pmddomingos This is quite likely correct, and smaller companies and countries would be wise to take it into account when planning AI, economic and social policies. @NZGovtFeeds
Both are valuable lines of research, and I don't see any reason to think that one is somehow stealing effort from the other (well, apart from the human cognitive biases that cause us to assume that comparable situations are zero sum games). What we have learned so far strikes me as both interesting and important.
A observation about safety culture as applied to programming. Not every risk can or should be prevented. Where possible, at reasonable expense and without imposing on people, risks should be mitigated. But what really needs to be done is prevention or redemption of substantive harms.
If you use its voice dictation mode, don't install ChatGPT Version 26.1027.30824 on the Mac. It's had an extreme regression in transcription accuracy, and is currently dangerously unusable. @OpenAIDevs@OpenAI
@JeffDean And we can expect the comparison to get ever better. One question though, is the distribution of environments for the two groups the same in these numbers?
@Plinz I've had exactly the same experience, and the same treatment, in my interactions. There are real, unresolved questions here. The way to answer them well for humanity goes through doing so together, not by staking out rigid positions in fortified camps.
@elonmusk it would be a fantastic customer incentive if @Starlink customers could connect on equipped planes irrespective of the frequent flyer program gates. Sent from StarLink on United
Why are we using the word "recursive" to describe tail recursion, or, in other words, a loop?
Related, what would "real" recursive self improvement loop like? It would have to involve subsidiary improvement processes that serve to enhance the ones above them.
@fchollet Agreed - either the singularity began when humans first invented technology (speech, fire, that sort of thing), or it begins when one passes the comprehension event horizon. It is reasonable at this point to think that, in Ray Kurtzweil's phrasing, the singularity is very near.
The entire argument, in the limit, for investment in vast data centres by AI companies is that the scale will enable them to deliver AI at lower cost than sovereign or on-prem models. There is a strong case for both, but that case is not cost saving.