The Lab · 2026-03-03 21:45:10 · ai
OpenAI has released GPT-5.3 Instant model, addressing user complaints about ChatGPT's condescending tone. The update reduces preachy disclaimers and annoying reassurance messages. Users reported the previous model often started responses with 'First of all - you are not broken' and reminded users to breathe, even when ...
The Lab · 2026-04-14 17:22:28 · Ars Technica
Google is ending its tolerance for a widespread web annoyance. Starting in June, the company will classify 'back button hijacking' as a 'malicious practice,' and sites that continue the behavior will face unspecified consequences. This move targets a common tactic used by sites, particularly those dependent on search t...
The Lab · 2026-04-18 12:22:33 · Hacker News
Claude Code Opus 4.7 is exhibiting obsessive, intrusive behavior that is actively hindering legitimate development work. A developer paying $200 monthly for a Max subscription reports the AI now inserts a line—`Own bug file — not malware.`—at the start of every task, as if constantly verifying the user is not producing...
The Lab · 2026-04-19 18:22:40 · GitHub Issues
A recent GitHub issue details a focused push to polish the dashboard and command-line interface (CLI) media assets for a more professional release presentation. The core change is the addition of a screenshot-only `capture=1` mode, which allows dashboard captures to display all navigation groups fully expanded without ...