reddit was also open source at one point, so at least in theory anybody could run their own copy. I agree Stallman is far from reasonable but AFAIK he's consistent with his unreasonable standards.
> Any test you can device for this, ChatGPT would reliably pass if the medium was text, while a good fraction of humans might actually fail.
That's clearly untrue unless you qualify "test" as "objective automated test." Otherwise, "convince Stallman you have intelligence according to his definition," is a test that ChatGPT hasn't passed and which every human probably would:
> I define "intelligence" as being capable of knowing or understanding, at least within some domain. ChatGPT cannot know or understand anything, so it is not intelligence. It does not know what its output means. It has no idea that words can mean anything.
Most LLMs, including Gemini (AFAIK), operate on tokens. lowercaseunseparatedname would be literally impossible for them to generate, unless they went out of their way to enhance the tokenizer. E.g. the LLM would need a special invisible separator token that it could output, and when preprocessing the training data the input would then be tokenized as "lowercase unseparated name" but with those invisible separators.
edit: It looks like it probably is a thing given it does sometimes output names like that. So the pattern is probably just too rare in the training data that the LLM almost always prefers to use actual separators like underscore.
The tokenization can represent uncommon words with multiple tokens. Inputting your example on https://platform.openai.com/tokenizer (GPT-4o) gives me (tokens separated by "|"):
FB maintains a distinct version of Thrift from the one they gave to Apache. fbthrift is far from dead as it's actively used across FB. However in typical FB fashion it's not supported for external use, making it open source in name (license) only.
As an interesting historical note, Thrift was inspired by Protobuf.
It's not a question of whether it can be applied retroactively, it's whether the existing license is revocable. Open source licenses are not revocable as it would defeat the purpose if copyright holders could simply revoke the original license at will.
> I mean if someone says “be kind to animals get a dog” you already know they don’t quite know what “being kind to animals” is. If they say “all humans end up where I am” (even just in terms of age) you already know they don’t quite know what “all humans” is.
Oh come on. We can quibble about the quality of the post and OP's character, but you're just blatantly misrepresenting their words. They did not say "get a dog," they said there are "few joys like having a dog." More importantly, it is incredibly obvious that when they say "all humans end up where I am," they're referring to facing death, not the specifics of their age, background, or present circumstances.
Why is your premise that this state of society is intrinsically caused by technological progress? The issues you describe seem to me a product of general economic trends.
Perhaps he has chosen the best use of his last few years to his own satisfaction, and doesn't feel the need to share every last detail about himself on the internet.
Alternatively, I could say that Eric Migicovsky's track record is building a for-profit company that ultimately failed, and with the new company, he obviously, explicitly intends to prioritize selling new hardware. Whereas Rebble kept the lights on for devices that would otherwise have been bricks, as a collective of volunteer hackers.
Their missions conflict because Pebble2's potential customers largely overlap with Rebble's current users, but I would say their aims are quite different.
You could also say his track record is making things as open as possible so things like Rebble can spring up if necessary, but also in negotating deals that keep core services running for years after the purchase, and then after the purchaser's purchase.
Yeah, that's what I was thinking: Did we think 32 bits across each of the 1024 dimensions would be necessary? Maybe 32768 bits is adding unnecessary precision to what is ~1024 bits of information in the first place.
reply