You got some shitty comments in here but you handled them well. While I think there is _some_ valid criticism, I think what you've built is pretty cool. I'd like to be able to test this without next JS. Is a CDN option planned for those wanting a quick start?
I enjoyed the read. How did you tap into the legacy Power Basic engine? Was there a FFI or some kind of bridge you could hook into? And what languages were you using?
I built an FFI via event dispatching and shared memory pointers/matching structs. Imported the C++ UI layer as a DLL via Win32 LoadLibrary. The PB shares a big array for storing global pointers allowing them to read/write each other's memory. The C++ has an event queue and has DLL functions the PB can call to peek/pop the next event. It actually isn't that complicated, just took me forever to come up with the idea.
As for languages, PB, C++, and JavaScript (Electron/Preact). I chose a no-build UI framework so that it could be modded by players without installing any build tools, just edit the text files in the game folder, and it has been a very good decision.
Indeed very clever. I wonder if you framed this problem up with claude how it would “guide” you to solve this problem. Would be an interesting match up of ai vs human. Love the story!
The feat was done before Claude Agent which is why it was so challenging. Although I admit I am a heavy user now circa the past two weeks. We shall not discuss my Claude Code experience lest I have another mental breakdown at work and my employer has to send me home again. Let me put it this way. I have set up Claude dangerously skip permissions with Agent Teams, Fast Mode, and our automated e2e test suite I designed where it can see screenshots of every step and browser and API console logs. It is entirely hands off software development. I have had to think long and hard about my identity as a software engineer. So forgive me if for my passion project I don't let Claude do everything, lest I remember the decades I spent reading those textbooks on my shelf, and the fear that I will forget it all.
Agree completely. Looking at OpenAIs public financials, Anthropic is likely exposing the true cost, where OpenAI is incinerating unlimited cash for users.
I think eventually companies will have the consumers foot the power bill somehow (local models?). No idea.
Same experience here. I started our devoutly using Claude but ran into some many limits that I switched back to ChatGPT and it's been night and day. I haven't even really been able to play with the Opus model on my Pro plan because it devours usage and then blocks me for X hours until it resets, costing me a work day. OpenAI has never done that to me. In fact, Codex just churned away for 2 hours on a task and I'm still using it without hitting a limit. I used to love using Claude but the limits are too prohibitive.
My experience trying to use Opus 4.5 on the Pro plan has been terrible. It blows up my usage very very fast. I avoid it altogether now. Yes, I know they warn about this, but it's comically fast how quickly it happens.
I feel like I can't even try this on the Pro plan because Anthropic has conditioned me to understand that even chatting lightly with the Opus model blows up usage and locks me out. So if I would normally use Sonnet 4.5 for a day's worth of work but I wake up and ask Opus a couple of questions, I might as well just forget about doing anything with Claude for the rest of the day lol. But so far I haven't had this issue with ChatGPT. Their 5.2 model (haven't tried 5.3) worked on something for 2 FREAKING HOURS and I still haven't run into any limits. So yeah, Opus is out for me now unfortunately. Hopefully they make the Sonnet model better though!
Exactly. I barely had a chance to kick the tires the couple of times I did this before it exploded my usage. I don’t just chat with it casually. The questions I asked were apart of an overall planning strategy which was never allowed to get off the ground on my tiny Pro plan.
Yeah same. Even though I find Opus-es to be more well-rounded (and more useful) for certain tasks, I instinctively reach for ChatGPT / codex to avoid burning up my usage limits for "trivial" work.
reply