I'd prefer a 30 minute response from GPT-5 over a 10 minute Response from {Claude/Google} <whatever their SOTA model is> (yes, even gemini 3)
Reason is: while these models look promising in benchmarks and seem very capable at an affordable price, I *strongly* felt that OpenAI models perform better most of the times. I had to cleanup Gemini mess or Claude mess after vibe coding too much. OpenAI models are just much more reliable with large scale tasks, organizing, chomping tasks one by one etc. That takes its time but the results are 100% worth it.
Hacky workarounds aren't rare exceptions; they're the plumbing of modern software. Anti-cheat and antivirus tools only work because they lean on strange kernel behaviors. Cloud platforms ship fixes that rely on undefined-but-stable quirks. Hardware drivers poke at the system in ways no official API ever planned for.
Yeah, they're ugly, but in practice the choice isn't between clean and hacky; it's between shipping and not shipping. Real-world software runs on constraints, not ideals.
On the other hand, everything you ship outside of a clearly established golden path is a maintenance burden that piles and piles and piles. And these maintenance burdens tend to gradually slow the org down until they cause rather catastrophic failures, usually out of security or hardware (read: fire) incidents. Or HR reasons because people figure there are better places to fight fires.
I ran "family guy themed cyberpunk 2077 ingame screenshot, peter griffin as main character, third person view, view of character from the back" on both nano banana pro and bfl flux 2 pro. The results were staggering. The google model aligned better with the cyberpunk ingame scene, flux was too "realistic"
i think they focus their dataset on photography. flux 1 dev one was never really great at artistic style, mostly locking you into a somewhat generic style. my little flux 2 pro testing does seem to verify that. but with lora ecosystem and enough time to fiddle flux 1 dev is probably still the best if you want creative stylistic results.
I looked through some of the GH repositories and - dear god - there are some crazy sensitive secrets in there. AWS Prod database credentials, various API keys (stripe, google, apple store, ...), passwords for databases, encryption keys, ssh keys, ...
I think hijacked NPM packages are just the tip of the ice berg.
> SPy is not a "compiler for Python". There are features of the Python language which will never be supported by SPy by design. Don't expect to compile Django or FastAPI with SPy.
Classic rug pull though