darby_eight

> its reasoning capabilities

To be clear, LLMs are not capable of reasoning.

> deny the risk entirely, or to act as if it's ridiculous to be concerned about such things in my opinion is just an intellectually ignorant position to hold.

This would take someone actually articulating what this risk you refer to is. All I have to go off of is Terminator, which is patently ridiculous. People in power would never relinquish this power to a computer without being able to dictate whom it benefits.

> The risk of the rise of AGI

can someone explain what the risk of this actually is? I just assumed it was a regulatory capture ploy.

IMO it is massively distracting from the very real and very measurable impact the global industry has on our climate.

> Every language has morphemes for the simple reason that every word is at least one morpheme.

Sure, if you define "morpheme" as a collection of syllables that's meaningful to people using alphabetic script. I don't see any benefit to this compared to working with syllables directly, which is a meaningful concept regardless of the script used to encode them.

> you still need to disassemble the ROM up front and annotate it heavily before it can be recompiled.

Why? Surely it's more straightforward to do binary-to-binary translation. No human input needed!

> Anyway, around here, even a small investment in reducing idling (or just cutting funding for traffic quiescence projects) would be equivalent to increasing vehicle fuel economy by something like 20-80%.

sure, but I'd rather just lobby against cars at that point. Especially in a place like the Bay Area that should be mostly served by public transit and dense housing. Lobbying to micromanage idling just feels like a waste of effort for such a negligible benefit.

ok, use a flaming trash can instead. Same result.

? Is this a joke

TBH I never thought to even call the speech processing in my a head a "little voice" until someone accused me of not having an inner monologue. I don't perceive it as audio at all unless I put effort into it.

Meanwhile music plays in my head constantly, so I don't think they're related.

Only in languages that have morphemes! This is hardly a universal attribute of language so much as an attribute of those that use an alphabet to encode sounds. It makes more sense to just bypass the encoding and directly consider the speech.

Besides, considering morphemes as semantic often results in a completely different meaning than we actually intend. We aren't trying to train a chatbot to speak in prefixes and suffixes, we're trying to train a chatbot to speak in natural language, even if it is encoded to latin script before output.

> On a character level this should be trivial.

Characters are not the semantic components of words—these are syllables. Generally speaking, anyway. I've got to imagine this approach would yield higher quality results than the roman alphabet. I'm curious if this could be tested by just looking at how LLMs handle English vs Chinese.

In the sense they're used as English prefixes, paleo is many orders of magnitude older than archeo. Hence words like "paleoarchaeology", which starts around 10kya.

Sure, but I'd be absolutely flabbergasted if they didn't come from kidnapped Palestinians.

> Graphics and PDFs are much lower on the priority scale.

Unless of course you read books that have graphics or come in pdf form.

I am not watching a video