lukev

38 ms·

511.
by lukev 3y ago
You are are literally describing a "lingua franca" here. And there's no evidence that English is special in it's ability serve as a lingua franca. In previous eras, other languages held this role (e.g, French, hence the
512.
by lukev 3y ago
Unless employees unionize.
513.
by lukev 3y ago
Everything you say applies to all large breeds. Judge a dog based on its owners and behavior. Don’t judge someone just because they own a pitbull. The breed isn’t the fundamental problem.
514.
by lukev 3y ago
I’ve had numerous rescues as well, over my life, and this tends to be generally true for all breeds. The defining feature of a rescue is that they were not trained and socialized well.
515.
by lukev 3y ago
There is absolutely nothing wrong with a well trained and socialized pit bull/mix. Anyone who says otherwise is lying to you or themselves. EDIT: None of the papers folks have posted control for training or socialization. I grant you t
516.
by lukev 3y ago
This one does/did.
517.
by lukev 3y ago
That's not how this place works, and clearly it would have been less offensive behavior if it was.
518.
by lukev 3y ago
No, in this case, it was a flat rate up front. Basically 100% a buffet except the table moves to you (which, to the point of the original article, waiting for your preferred sushi to come around probably slows down consumption and improves
519.
by lukev 3y ago
This place charged a flat fee at the door, then everything inside is all-you-can-eat.
520.
by lukev 3y ago
I love articles like this. I’d happily read a book with each chapter breaking down the unit economics of a given type of business. Also, I can’t help but recount an anecdote from the last time I was at an all-you-can-eat place… conveyor-bel
521.
by lukev 3y ago
I'm a little surprised this is newsworthy... this is already the policy at my kid's middle school (NC public schools.) Is it that uncommon? I get that it's probably harder to enforce into the upper grades...
522.
by lukev 3y ago
Interesting points. But I do question the assertion that humans are innately more willing to trust natural language interfaces. Humans are evolutionarily primed both to trust and to distrust other humans depending on context. It might act
523.
by lukev 3y ago
yes, except the "reasoning" process should also be able to look up facts (retrieval) and invoke external tools, making it non-pure.
524.
by lukev 3y ago
Completely agree. From a computer science point of view: a single prompt/response cycle from a LLM is equivalent to a pure function; the answer is a function of the prompt and the model weights and is fundamentally reducible to solving
525.
by lukev 3y ago
Not without extra machinery around the model itself. Token generation is always probabilistic to some degree.
526.
by lukev 3y ago
Wonder how long before Qanon conspiracy nuts latch on to this naming pattern. Sigh.
527.
by lukev 3y ago
There's actually a somewhat reasonable analogy to human cognitive processes here, I think, in the sense that humans tend to form concepts defined by their connectivity to other concepts (c.f. Ferdinand de Saussure & structuralism).
528.
by lukev 3y ago
> In addition hot people are just given more latitude than ugly people in terms of social faux pas. No no, that's exactly the point. If an advance is welcome, it's wasn't a faux pas. If it's not welcome, it's a f
529.
by lukev 3y ago
But that's not helpful for driving any kind of informed opinion or public health policy. It'd be like classifying nearly all deaths as caused by "cardiac arrest" even though that's the natural end state of most othe
530.
by lukev 3y ago
Regardless of one's opinions on if MAID is a good idea, it seems unfair to regard it as a "cause of death". Wouldn't that be more correctly attributable to the underlying condition?
531.
by lukev 3y ago
This notion (step 1: be attractive, step 2: don't be unattractive) comes up a lot in male-dominated online spaces. It's a fallacy, and actually a self defeating one since it causes guys to ascribe their failures to something out o
532.
by lukev 3y ago
Yes, that's fundamentally it. By making the marking more liquid, different types of entities can adjust for their specific predictability/risk needs. Also note that though these markets are "free" they are still extremel
533.
by lukev 3y ago
Genuinely, great distinction. Would you also include in “AI ethics” a considered rollout of automation to maximize its benefit to everyone ? Because there is real benefit. The kind of work AI can automate is shit work and it would be to th
534.
by lukev 3y ago
If you read the article, that’s literally what this is talking about. There’s a simple diagram and everything.
535.
by lukev 3y ago
The statement says. It would mean not just a misalignment on values but active deception regarding OpenAIs current direction. The bit about “ability to fulfill duties” sticks out, considering the responsibility and duties of the nonprofit b
536.
by lukev 3y ago
The board in question is the non-profit board. If Sam was pursuing profits or growth (even doing a really good job of it) in a way that violated the objectives set by the non-profit board, that could set up this kind of situation.
537.
by lukev 3y ago
Unless Altman was taking actions in furtherance of the for-profit goals, while abandoning the non-profit goals, and not being honest to the board about this. This actually seems the most probable reason for this given the circumstances and
538.
by lukev 3y ago
Using retrieval to look up a fact and then citing that fact in response to a query (with attribution) is absolutely within the capabilities of current LLMs. LLMs "hallucinate" all the time when pulling data from their weights (bec
539.
by lukev 3y ago
A body may be a necessary condition for epistemological knowledge, but it's certainly not sufficient... most animals are embodied but do not possess propositional knowledge (IMO anyway.) Ultimately though this is all about the definiti
540.
by lukev 3y ago
I mean, this is why philosophical skepticism is a thing. There's hundreds of books written on epistemology. But I think almost none of them would ascribe "knowledge" to a statistical engine that output propositions of random

More ›