Ask HN: Is it AGI for software engineering? | Hacker News Reader

Ask HN: Is it AGI for software engineering?

3 pointscolesantiago5 comments

With both Opus 4.6 and GPT 5.3 Codex out, it is now clear that software engineering is probably beyond solved as AGI might come before 2030.

The last pieces are being solved progressively and quickly at each model release for example with Security [1]

It is only a matter of time before, Infrastructure, Mobile, Maintenance, etc. will just be a solved problem.

[1] https://www.axios.com/2026/02/05/anthropic-claude-opus-46-software-hunting

View on HN

I feel like the models are already smart enough to be called something like an early version of AGI. It's the tooling around it that needs to improve.

For example, Codex/Claude should be able to code and "see" its changes. If it's developing a mobile app, it should be able to view the real app it's building and then make changes. If it's a website, it should see the changes being rendered and then make further changes if needed.

The tooling is well behind model intelligence.

DBs are a solved problem for most cases. APIs are a solved problem. UX patterns are a solved problem. Yet I solved problems related to these things every day.

Isn’t AGI described in contrast to specific intelligences? How can you have specific AGI?

    it is now clear that software engineering is probably beyond its own AGI

Clear to who? Try engaging your brain instead of parroting marketing materials

My question is more clearer, english isn’t my first language.

Reply on news.ycombinator.com