Gotta admit I'm still trying to get the hang of building with AI. It's amazing at creating showcase stuff. "Make me a terrain." <makes a nice terrain> "Make me a game about foo." <makes a game about foo> But iterating on something with an actual idea in mind? It trips over itself non-stop. Implements one feature by destroying 3 others, etc... I've been trying to rein it in torwards making an actual "game" and not just a demo, and I'm literally just spelling out all the systems it's supposed to make. me: "Did you using the loading API we made?" ai: *thinks* Nope, sorry. My bad. Even though I said I used it, I only actually used it in one case, hard-coded random loads in 4 others, and literally compiled the assets into the executable in these other cases. me: wtf Am I doing this wrong, or is that anyone else's experience?

Open Al's internal benchmarks show GPT-6.1 Sol crushing Opus 5.5, with Anthropic struggling to keep up

Idea for Netflix… My girls (aged 5 & 7) wanted to watch Legally Blonde recently, so we watched it even though it’s rated 12. There’s only a bit of swearing and one or two scenes that aren’t really suitable for them. We ended up fast-forwarding through one scene and basically hoping they didn’t pick up on the bad language. It got me thinking - how useful would it be if Netflix had an option to automatically bleep/mute bad language and skip or mute certain scenes? I’d happily watch a slightly “cleaned up” version of a film with them rather than having to avoid it altogether. There are so many great films rated 12 or 15 that are only unsuitable because of some swearing and the odd dodgy scene. Seems a shame to have to wait years to watch them. Surely this would be technically possible? It would be a genuinely useful feature for families. cc @netflix

i’ve been wanting to make this edit since i was 8 and i finally did now this is almost exactly how it’s been in my head for 12 years