AI - AI Dynamics

→ View original post on X — @reach_vb

GPT 5.5 Launches Fast Mode in Codex for 1.5x Speed

By

@reach_vb

–

29 April 2026 7h57

PSA: GPT 5.5 ships with /fast mode in Codex Same model. 1.5x FASTER

→ View original post on X — @askalphaxiv

Kwai Summary Attention: Learnable Tokens for Long Context

By

@askalphaxiv

–

29 April 2026 7h56

"Kwai Summary Attention Technical Report" This paper uses learnable summary tokens for long-context attention. It splits text into chunks, compresses each chunk into a summary token, keeps recent text dense within a sliding local window, and reads distant context through

MIT study questions AI automation ROI amid massive capex spending

By

–

29 April 2026 7h55

MIT finding AI automation only economically viable in 23% of vision-primary roles sits in direct tension with $740B in capex and 92,000 layoffs. Those two things cannot both be rational at the same time.

AI Transition: Jobs, Spending, and Economic Uncertainty Ahead

By

–

29 April 2026 7h53

The optimist answer is new jobs, new spending categories, wealth redistribution. The honest answer is nobody knows if that transition happens fast enough or broadly enough this time.

Clean Architecture Improves Product Velocity in Software Development

By

–

29 April 2026 7h51

Clean architecture matters more than users realize until the old one is gone. If this ships well, the product velocity on X should noticeably improve within a few quarters.

GPT-5.5 Spud shipped, monthly LLM cadence prediction confirmed

By

–

29 April 2026 7h50

Already happened. Spud shipped as GPT-5.5 six days ago, one week after Opus 4.7. The monthly cadence prediction was right, the timeline was just tighter than expected.