shwaj

9 ms·

121.
by shwaj 1y ago
At some point it becomes difficult to improve the O() complexity. How do you do better that the O(n-squared) of the Transformer, with acceptable tradeoffs? Many big brains in all the big labs are very aware of the importance of algorithmi
122.
by shwaj 1y ago
I’m not a practitioner. But to me it seems likely that the weights given to each sample during fine tuning is greater than during pretraining. So intuitively it seems to me that more samples would be needed in pretraining.
123.
by shwaj 1y ago
Fuchsia seems like a potentially practical OS for constraining the operations of AI models. As an object capability operating system, each component (and hence the process that one is instantiated in) has access only to the capabilities tha
124.
by shwaj 1y ago
Are you sure? It wouldn’t shock me, but they specifically say “Gemini Apps”. I wasn’t familiar with the term, but a web search indicated that it has a specific meaning, and it doesn’t seem to me like web search AI summaries would be cover
125.
by shwaj 1y ago
> Today, I consider anyone who isn’t an emergency responder with a phone that makes noise a selfish asshole. You’re probably getting a lot of false positives. Is it ok to answer a vibrating phone around you, or do I need to leave the vi
126.
by shwaj 1y ago
You might want to look into the work of Michael Levin’s lab to give you a broader perspective on the intelligence of individual cells and tissues. Your machine metaphor is arguably as misleading as anthropomorphization. Edit: did you read t
127.
by shwaj 1y ago
In this context I believe it refers to models that are trained to generate an internal dialogue that is then fed back in as additional input. This cycle might be performed several times before generating the final output text. This is in c
128.
by shwaj 1y ago
I think it’s a play on “don’t anthropomorphize the lawn mower”, referring to Larry Ellison.
129.
by shwaj 1y ago
The Internet and the Web are not the same thing. Apple absolutely gatekeeps the ability to run non-HTTP internet connected apps. Your statement that Apple is not a “gatekeeper to any part of the Internet” is simply false.
130.
by shwaj 1y ago
So smug. Unless you’re replying to someone you know (and my apologies if so), how do you know their die hard libertarian friends voted for Trump?
131.
by shwaj 1y ago
I think they were replying to the person with the green user name :-)
132.
by shwaj 1y ago
It’s true that the large majority of aid came in the form of weapons that the military industrial complex wanted to replace anyway. However, there has still been many billions of “cash” sent, totaling over $40 billion. That’s nothing to sn
133.
by shwaj 1y ago
People who disagree with the Covid response may have been voting for RFK as much as Trump.
134.
by shwaj 1y ago
Many people disagree with the Covid response. Many people look at the Bidens’ relationship with Ukraine and wonder whether the war could have been avoided, and we wouldn’t have to hear stories about “we sent $X billions, but Zelenskyy says
135.
by shwaj 2y ago
It’s somewhat surprising to me. If you take any two furry animals, they are more closely related to each other than to a lizard.
136.
by shwaj 2y ago
How would that work, exactly? Would Google’s search algorithm highlight pro-TRUMP articles to pump the token price? I think you’ll be waiting a long time.
137.
by shwaj 2y ago
Or perhaps if your phone goes down.
138.
by shwaj 2y ago
Makes sense, thanks. I wonder whether human web-browsing strategies are optimal for use in a LLM, e.g. given how much faster LLMs are at reading the webpages they find, compared to humans? Regardless, it does seem likely that Google’s dat
139.
by shwaj 2y ago
Can you say more about using RL at inference time, ideally with a pointer to read more about it? This doesn’t fit into my mental model, in a couple of ways. The main way is right in the name: “learning” isn’t something that happens at inf
140.
by shwaj 2y ago
I love how you respond seriously to the logistical challenges of salt water, and don’t blink at the prospect of using nukes to create a tidal wave.
141.
by shwaj 2y ago
It’s unfortunate that Claude can’t provide references, but it plausibly suggests that the indigenous population actively managed forests in a radius around their population centers, not the entire land area. So the current larger population
142.
by shwaj 2y ago
“Filters are delays” doesn’t imply “delays are filters”. In particular, the type of effect known as a delay (e.g. a delay guitar pedal) isn’t a filter, certainly not in the musical sense, which is the relevant sense here.
143.
by shwaj 2y ago
Me too. White screen on Chrome, iOS and macOS.
144.
by shwaj 2y ago
By their definition, an amplitude envelope would probably also be a transformation.
145.
by shwaj 2y ago
Delays aren’t filters.
146.
by shwaj 2y ago
Sorry if I wasn’t clear. I meant that elite military forces make peace with death as part of their training. So, “having a lot to lose” would be less of a deterrent compared to those without such training.
147.
by shwaj 2y ago
The “new generation of paramilitary leaders” described by the article includes “doctors, career cops and government attorneys”. All of these would seem to have a lot to lose. If this is true, one would think that elite military operators
148.
by shwaj 2y ago
Furthermore, the journalist explicitly encouraged suspicion toward Williams: > Remarkably, the AP3er defended Williams’ loyalty each time I asserted he’d secretly tried to undermine the group. Is this how journalists protect sources? Se
149.
by shwaj 2y ago
I agree completely, and yet I would still prefer language to be used consistently.
150.
by shwaj 2y ago
I know it’s not necessarily the same people, but it feels contradictory for this community to say “copyright infringement isn’t theft” when we’re talking about movies, but use the opposite language when talking about GPL source code.

More ›