What would be an interesting LLM announcement, then? Even Opus 5.5 wasn't an interesting announcement but the impact is obviously significant in hindsight.
All this incremental progress is just more of the same. I want a real novel capability, not just more, faster, cheaper, “smarter”.
We are barely 4 years in since the release of a historically important tech. Of course the next few years are incremental. thats how it always works.
Even last year, before opus 4.5, long-horizon agents didnt really work. I would not say the last 12 months are only incremental
As an aside, I remember when sora 2 launched and thinking, "wow I wonder if we'll ever get something close this in open weights?" The prevailing sentiment was "maybe 5 years." Yet, we have Minimax H3 and Flux a year and a half later.
But this only works for frontier models. If MiniMax catches up to Opus 5.5 in 6 months, no one would notice.
I'd be happy as hell if I end up being able to run an Opus 5.5-level model at home within a 6-month time horizon. May or may not happen based on China's export policy, of course.
The open weights we have may be as good as they ever get... but then that's been true all along, considering nobody really knows why they are releasing powerful models for free in the first place.
I remember that close to an year ago i felt the same way about opus 4.5. And even if you can easily run ds or qwen locally to get very similar results to that, we moved to 5.5 being the new target.
Deepseek can do 90% of the job but a claude reviwer/spec writer is a brutal improvement.