I'm even more pleased with this because basically everything I tried worked first time ;) no silly errors with my maths - turns out three years of sleep deprivation haven't entirely killed my ability to do high school level maths

Thank you! There are many issues, mostly down to me spending less than 10 minutes on enemy spawning 🫠 and I've essentially spent my life training for this exact moment - I've done somewhere in the region of 50 jams that lasted an hour or less. This was almost *easy* in comparison!

it is worse even than that! If you die without scoring, you can continue to play in the background 😂 and then if you die again, the ui freaks out a little. Admittedly it does then get back to the proper state, but if you really want to you can keep playing 😉

and here is theirs. I hate this and everything it stands for 👍
misterm.itch.io/bloodbullets...

Basically, after reading your whole post I remain unconvinced that game loops are the problem, sorry 😂

Without a good core loop, you'd definitely finish fewer games. Runtime is the bigger issue, and AAA risk adverse slop that doesn't really have any meaning and takes no risks. Plus we're in a market where players value $ per hour of play time, which as a metric just suuuucks.

Movies are a couple of hours, you do the whole thing in one sitting. Books are longer than movies, but also less demanding and are _still_ significantly shorter than games. Most games target 20+ hours and you need to be actively engaged and participating throughout!

It's well written and an interesting read! I find the premise of "I finish movies and books but not games, isn't that interesting" a tad frustrating because there is a huge difference in scale 😉

Yeah Pokémon mystery dungeon did something like this, you can pick them up and pay by talking to kecleon, or attempt to steal and run away 😂 I think you could also drop stuff to sell it?

Also this, from a study by stack overflow! Not game development specific, but it tracks with my general experience - more bugs, zero improvement to delivery time.

I just see it as "brain off" programming, which I guess is fine sometimes - especially when you've got a game to release. But it isn't changing the industry in a positive way

In my experience, code produced by LLMs is slower, more prone to errors, and less understandable for anyone working on it. If AI gives you a 10x productivity boost, you're about as productive as 0.1x a normal programmer 😉

Using LLMs to help code basically turns you into a code reviewer, not a programmer. People aren't good at reviewing code by default! You will miss things because it looks "good enough". Also you're no longer solving problems, you're handholding a machine that does not understand the problem space.

Hey nice one! 100 days of dev REALLY helped me get my game over the line, I wrote a blog post every day, with a few saved up for breaks. Good luck! Have you got a link to the jam project? I can give some feedback if that would help!

Not a problem at all, thanks for the response!

This looks awesome! You don't happen to take paid commissions do you? 🤩

It also gives you a really nice way to do demo builds that only include specific scenes, or split builds for stage & prod so you can preview changes in a test environment. Lovely stuff

Would love to be added - self published on Xbox!