me: please find security issues
fable 5: Fable 5's safety measures flagged this message for cybersecurity or biology topics. They may flag safe, normal content as well. These measures let us bring you Mythos-level capability in other areas sooner, and weβre working to refine them. Switched to Opus 4.8
ewβ¦ π₯΅
Just had a very nice session with opus-4.8-high regarding some of my humble ML work, where we made substantial progress on top of the current state I achieved with gpt-5.5-xhigh. Sadly, we only tested and invalidated a bunch of ideas, did not get any breakthrough, but I feel like I understand more of the stuff and the approach of opus-4.8 is more systematic and less mechanistic than gpt-5.5
doing apartment renovation in 2026: openclaw for everything
materials selection, checking bills, contacting contractors, how-to-accept the results of each stage and so on and so on
@i_Kisliy macOS / iOS search is pure chaos
I literally type in the full name of the app into Spotlight and get "No results". Has been like that for years now
court-vibing: we just filed a court motion with openclaw, opus 4.7 and gpt-5.5 pro
- arbitrage court for a bankruptcy case
- github repo for openclaw agent
- ingestion pipeline for all court documents (pdf->md, daily cron + manual)
- opus 4.7 as a daily driver
- gpt-5.5 pro as a reviewer (really helped)
the assigned bankruptcy trustee is asking: who is that person who can spot such tiny details (it is my claw lol)
we'll see how it goes
court-vibing: we just filed a court motion with openclaw, opus 4.7 and gpt-5.5 pro
- arbitrage court for a bankruptcy case
- github repo for openclaw agent
- ingestion pipeline for all court documents (pdf->md, daily cron + manual)
- opus 4.7 as a daily driver
- gpt-5.5 pro as a reviewer (really helped)
the assigned bankruptcy trustee is asking: who is that person who can spot such tiny details (it is my claw lol)
we'll see how it goes
Built a basic Symphony like orchestrator today. Talked to gpt-5.5 in plan mode where it bombarded me with pretty thoughtful questions for 30 minutes
Then it built it in 10. With a few rounds of bug fixing (acli permissions / jira stuff π) it got it running and the first ticket got through:
- I assigned a label to the ticket
- orchestrator running locally picked it up (acli polling)
- built to completion, ran checks
- made a PR
second day in a row gpt-5.5 IQmogging opus-4.7
yesterday it was proposing very nice and compact changes based on PR reviews, better than what I saw recently with opus-4.7
today gpt-5.5 xhigh re-implemented a PR from scratch and comparing it to previous opus-4.7 xhigh implementation, the gpt-5.5 output is just better