Something weird happened with this post. It appeared in a shadow banned sort of form (maybe because of the Unicode superscript?) so I made the other one instead, but now this one actually appeared. Mods should feel free to delete this; I don't think I can do so myself.
> because GIMPS always at least eventually gives out numbers to multiple people to check
For the 40th Mersenne prime, 2²⁰⁹⁹⁶⁰¹¹ − 1, for example, the status page on mersenne.org seems to suggest only a single check (and a handful of later NF checks later), but maybe follow-up proofs and proof certifications and reruns are omitted? https://www.mersenne.org/report_exponent/?exp_lo=20996011&fu...
Also, when you sign up, you're asked to provide an email address in case they want to get in touch with you, so even if OP didn't themselves do so, I imagine they would include their work just the same?
> However, your name could have ended up on that Wikipedia list as a discoverer. :-)
Along the same lines, since each (potential) prime is being worked on by many computers, some looking for factors, some running the Fermat test, some running proof certification work, who gets the "discoverer" title; just the person who ran the PRP test? If so, seems fair enough, since that's where most of the computational budget ends up, but on the other hand, it seems like that would disincentivize running anything but PRP tests. But maybe the people involved are just in it for the mission (or the GHz day leaderboards).
There's also https://scirate.com/ which occasionally has active discussions but, at least in my field, there's far from critical mass, and discussions only happen when someone kick starts and advertises a thread.
I believe the most active field is quantum information, which has enough activity such that paper get dozens of upvotes, but the conversation level is basically as you describe.
One I remember stumbling upon some years back was the .NET Framework equivalent of that, System.Random.Next(0, int.MaxValue), which has a much greater probability of producing odd numbers than even numbers, the probability of getting odd numbers being 50.34%, because of some rather unfortunate translations between integers and floating points. https://fuglede.dk/en/blog/bias-in-net-rng/
Yeah, sorry, that could have been clearer, I added a few more instructions. Basically, chances are that even if you've got Harfbuzz running, you're still running a version with no Wasm runtime. If so, chances are you can get away with building it with Wasm support, then add the built library to LD_PRELOAD before running the editor.
That was useful. I have indeed compiled and installed wasm-micro and now meson build it successfully. Tho "meson compile -C build" returns an error about not finding "hb-wasm-api-list.hh". Do you have any experience of that?
EDIT: Nevermind. Using the exact commits you linked give another error (undefined reference to wasm_externref_ref2obj). I give up
You got it, same seed in practice, but also just temperature = 0 for the demo actually. A few things I considered adding for the fun of it were 1) a way to specify a seed in the input text, 2) a way to using a symbol to say "I didn't like that token, try to generate another one", so you could do, say, "!" to generate tokens, "?" to replace the last generated token. So you would end up typing things like
"Once upon a time!!!!!!!!!!!!!!!!!!!!!!!!!!!!!SEED42!!!!!??!!!??!"
and 3) actually just allow you to override the suggestions by typing what letters on your own, to be used in future inferences. At that point it'd be a fairly generic auto-complete kind of thing.
Using the input characters to affect the token selection would increase the ‘magic’ a little.
As it is, if you go back into a string of !!!!!!!!!! That has been turned into ‘upon a time’, and try to delete the ‘a’, you’ll just be deleting an ! And the string will turn into ‘once upon a tim’.
If you could just keyboard mash to pass entropy to the token sampler, deleting a specific character would alter the generation from that point onwards.
This is fantastic! These days, we have learned that no product can be taken seriously if it doesn't use LLMs somehow, so also based on Harfbuzz-Wasm, here's a font which is also an LLM: https://news.ycombinator.com/item?id=40766791
