News for hackers. Stuff that matters.
Re:Grok 4.6 by causal on Wednesday August 12, 2026 @12:02PM (#49274571)
Agreed, but my suspicion is tied to the timing. Catching up eventually is to be expected. Having similar jumps in capability ready at the same time is odd.
Re:Grok 4.6 by dash2 on Wednesday August 12, 2026 @12:11PM (#49274693)
Maybe "readiness" is quite a flexible category? You're mid-training for your next model; a rival releases something; you clear the boards and release the model without completing the training run?
Re:Grok 4.6 by causal on Wednesday August 12, 2026 @12:15PM (#49274772)
Touche, aborted training runs probably do happen often. Closed model providers have zero incentive to announce a new model with less-than-best benchmarks.
Re:Grok 4.6 by noddybear on Wednesday August 12, 2026 @01:15PM (#49275696)
I don’t think the runs need to be aborted… you can just release a mid-training checkpoint!
Re:Grok 4.6 by verdverm on Wednesday August 12, 2026 @03:59PM (#49277788)
you'd be crazy to not be taking snapshots on the regular, many good reasons besides failures
Re:Grok 4.6 by ben_w on Wednesday August 12, 2026 @02:28PM (#49276676)
Gradual improvements in performance can look like jumps, when you go over critical thresholds.

Combustion engines improved gradually, each year. One year they got better than horses.

Re:Grok 4.6 by martinald on Wednesday August 12, 2026 @03:52PM (#49277695)
There's also a bit of selection bias going on here because we forget about labs that don't have a jump and just focus on the ones that do. Notably Google is definitely not having that capability jump.