cs702

52 ms·

751.
by cs702 2y ago
Thank you for that. It made me chuckle! Here are some counter-examples: * A new car company making a new kind of vehicle. The first time I heard he wanted to build a new car company that would sell EVs, I dismissed it as crazy. Tesla now se
752.
by cs702 2y ago
WTF? This is utterly bananas. I don't get it. The long-term cost of treating longtime, loyal, talented employees like Rebecca Tinucci as... disposable could be very high. It could even threaten the company's survival. But... I&#x2
753.
by cs702 2y ago
Wow, Rebecca Tinucci (EV charging) and Daniel Ho (EV products) are gone. WTF? This is utterly bananas. I don't get it. The long-term cost of treating longtime, loyal, talented employees as... disposable could be very high. It could eve
754.
by cs702 2y ago
Original source: https://www.theinformation.com/articles/musk-plans-more-layo...
755.
by cs702 2y ago
> I understand that's the context, but I'm not sure that it's unfair nitpicking. The OP is about much more than that, and taken as a whole, suggests the author is well aware that human beings absorb a lot more data from
756.
by cs702 2y ago
I agree with you, but your comment strikes me as unfair nitpicking , because the OP is referring to information that has been encoded in words.
757.
What Students Read Before They Protest(nytimes.com)
4 points by cs702 2y ago|2 comments
758.
by cs702 2y ago
> You can always represent removed state by additional state that represents deletion of whatever preceding state was there. Good point. Thank you!
759.
by cs702 2y ago
No. RAG is about finding relevant documents/paragraphs (via KNN lookups of their embeddings) and then inserting those documents/paragraphs into the input context , as sequences of input tokens. What I'm talking about is diff
760.
by cs702 2y ago
Maybe there was something on the water they drank as kids? (I'm joking, in case it's not obvious!)
761.
by cs702 2y ago
This is exactly the sort of gem I love to find occasionally on HN. Done not for money, but, if I may paraphrase George Mallory, because the challenge was there. Someone had to do it.
762.
by cs702 2y ago
Wow. The same Gary Kildall who was first approached by IBM for a PC operating system[a] helped write Pixar's first 3-D renderer. At a minimum, Kildall's Wikipedia page should be updated with his contribution to Pixar! --- [a] Bill
763.
by cs702 2y ago
The OP is not trying to answer the question. Rather, the OP is asking the question and sharing some thoughts on the motivations for asking it.
764.
by cs702 2y ago
...by Cory Doctorov. If you've read anything by him before, you will already have a sense of what he writes in this opinion piece. As always, he's very much in favor of unionization by tech workers, including engineers.
765.
by cs702 2y ago
Thank you for sharing this here. Rigorous work on the "expressibility" of current LLMs (i.e., which classes of problems can they tackle?) is surely more important , but I suspect it will go over head of most HN readers, many of w
766.
by cs702 2y ago
As an aside, at one point I experimented a little with transformers that had access to external memory searchable via KNN lookups https://github.com/lucidrains/memorizing-transformers-pytorc... ( great work by lucidra
767.
by cs702 2y ago
Fantastic essay. Highly recommended! I agree with all key points: * There are problems that are easy for human beings but hard for current LLMs (and maybe impossible for them; no one knows). Examples include playing Wordle and predicting
768.
by cs702 2y ago
> Perhaps by having a dataset of symbols that corresponds to symbols that you similarly understand how they can be put together, as well by having a model for how to respond to inputs given some prompts. That's the same as saying th
769.
by cs702 2y ago
For a machine to be able to say "what you want to hear," it must understand what you're talking about, internally, in some fashion. How else could it do that ? As Hans Moravec wrote in 1998, when he predicted intelligent m
770.
by cs702 2y ago
I disagree with the authors's key claims: 1. They claim that "an LLM does not understand the semantic meaning of a sentence in a linguistic sense, but rather calculates mathematically what the most likely next word should be based
771.
by cs702 2y ago
The key facts, for me: > Since it was first identified, the H5 strain of avian influenza and its variants have led to the slaughter of over half a billion farmed birds. Wild-bird deaths are estimated in the millions, with around 600,000
772.
by cs702 2y ago
I've tried this for Wordnet hierarchical classification with a Standard Cross Entropy loss: https://github.com/glassroom/heinsen_tree#sample-usage-with-... It worked for me, but I had to modify the code to use all
773.
by cs702 2y ago
TL;DR: The authors pretrain the model to classify images into Wordnet synsets[a] that appear in the caption, using a standard Cross Entropy loss. They keep the number of classes relatively small by removing any synsets that don't show
774.
by cs702 2y ago
Thanks. Yes, you're right, sometimes "treadmill" is used pejoratively to describe forced upgrades that otherwise would be unnecessary. I was referring to the fact that we are all voluntarily installing security updates an
775.
by cs702 3y ago
Good point! I'll edit my comment.
776.
by cs702 3y ago
Sorry, I don't understand your question. By "treadmill" I mean sticking to a particular distro year after year, upgrading regularly -- it could be Ubuntu, Debian, Red Hat, CentOS, etc. Unless you're constantly switching
777.
by cs702 3y ago
Looks like a no-brainer if you're on the Ubuntu treadmill. Key differences versus the previous LTS version: * Distribution packages built with additional security-hardening features * Linux Kernel 6.8 * Gnome 46 * Updated utils and app
778.
by cs702 3y ago
> an exercise of pure marketing Yes. Great choice of words. A lot of non -frontier models look like "an exercise of pure marketing" to me. Still, I fail to see the rationale for telling the world, "Look at us! We can
779.
by cs702 3y ago
> "build something fast, get users, engagement, and venture capital, hope you can grow fast enough to still be around after the Great AI cull" Snowflake is a publicly traded company with a market cap of $50B and $4B of cash i
780.
by cs702 3y ago
You're right. I changed it. Thanks!

More ›