If your application consists of a single service, there's virtually no point in building it. Anything that simple could be recreated with AI in a matter of minutes.
Now that the Atlas enterprise platform is getting to work, the research version gets one last run in the sun. Our engineers made one final push to test the limits of full-body control and mobility, with help from the RAI Institute.
I want so badly to use Grok 4.1 fast reasoning instead of Opus 4.6 or GPT 5.3 high, but it's just not feasible for anything really complex at this point. I genuinely enjoy using XAI, but Grok 5 better come soon.
If @netflix was smart, they’d make the next season of Night Agent about the current state of the government. An action show about a hero fighting corrupt pedos in the government would be massively successful and insanely entertaining.
“Software engineering is dead” only applies to those in government, health care, and other industries scared of innovation. The rest of the engineers that have worked with AI for years already will take all those jobs — one way or another.
I started a new job two weeks ago exactly. So far, I’ve submitted two PRs:
1. 150k lines added (a lot from moving an existing service into a monorepo)
2. 80k lines added (~40k being static assets or generated files)
Regardless, easily 50-60k lines of real added code.
The only people saying software engineering isn't over are the people using composer and agent mode for everything.
Start using plan mode, debug mode, Opus 4.6, skills, rule files, etc and tell me the same thing (you won't).