glial

18 ms·

211.
by glial 2y ago
ChatGPT currently replies: > It seems like there might have been an issue with your message. Could you please clarify or provide more details on what you need help with?
212.
by glial 2y ago
The article doesn't mention this, but I imagine having an AWS version of the supercomputer would be extremely helpful for software development and performance optimization, especially if the same code could be tested with fewer nodes.
213.
by glial 3y ago
Perhaps you aren't the target market for this.
214.
by glial 3y ago
IMO reproducing findings could/should be a mandatory part of PhD training.
215.
by glial 3y ago
The Chrome Web Store page doesn't say anything about annotations.
216.
by glial 3y ago
Economic "theories" have an annoying way of being self-fulfilling.
217.
by glial 3y ago
This is all true, but we need to start somewhere.
218.
by glial 3y ago
Also, https://jan.ai is open source and worth trying out too.
219.
by glial 3y ago
I think of COT as a memory scratchpad. It gives the LLM some limited write-only working memory that it can use for simple computations (or associations, in its case). Now suppose an LLM had re-writeable memory... I think every prompt-hack,
220.
by glial 3y ago
> corporations are still composed of humans Yes, but humans who use tools to efficiently and achieve their narrow set of goals, sometimes ruthlessly at the expense of others. Even if AI ends up only being a tool to increase this efficien
221.
by glial 3y ago
AI should replace lawyers, since they are essentially a regressive tax on engaging with the legal system.
222.
by glial 3y ago
True; I was thinking about abstract algebra, set theory, and number theory when I wrote that.
223.
by glial 3y ago
This is well-written. I might suggest that what makes pure mathematics special is that abstractions in pure math are not leaky, unlike in (nearly?) every other domain.
224.
by glial 3y ago
Comparability across test years is taken seriously by the folks that make these tests (see e.g. [1,2]). There are valid criticisms to be made, but flippantly calling the test 'rigged' is not constructive. [1] https://ww
225.
by glial 3y ago
Shorter is good, but dynamic tests (called "computer adaptive tests" or CAT in the industry) make it difficult to follow longstanding best-practices for test taking. When I took the SAT, I could quickly scan through the questions
226.
by glial 3y ago
If you're running LLMs locally using something like LM Studio or Jan.ai, a GPU bump will speed up text production.
227.
by glial 3y ago
I thought I hated JIRA until I moved to other companies that don't use it. One company allowed each team to manage tasks however they liked - and therefore each team had a different method. Other companies tried to force Trello or Asan
228.
by glial 3y ago
This is cool, but I'm amused that CatBoost still beats it.
229.
by glial 3y ago
> imagine they actually do produce a breakthrough LLM but don't know how to monetize it yet and traffic to google search craters ChatGPT has replaced maybe 1/2 of my Google searches and the cognitive relief from not having to w
230.
by glial 3y ago
There were 255 traffic deaths in NYC 2022[0], so according to the statistic on the site you link to, driving is 36x more dangerous. [0] https://www.silive.com/news/2023/01/nyc-traffic-deaths-drop-... .
231.
by glial 3y ago
Does anyone know whether this applies to clothing that uses recycled plastic? I know Patagonia uses lots of recycled polyester: https://www.patagonia.com/shop/collections/recycled-polyeste...
232.
by glial 3y ago
I always hear this caveat, and I'm sure you're right, but I'm curious about the threat model here. If I roll my own RSA using e.g. GMP[1], what would it take to find vulnerabilities? Is there some automated software that coul
233.
by glial 3y ago
> EVERYTHING MUST FAIL NORMAL This is the Mitch Hedberg Rule. "An escalator can never break: it can only become stairs. You should never see an Escalator Temporarily Out Of Order sign, just Escalator Temporarily Stairs. Sorry for t
234.
by glial 3y ago
To channel Cormac McCarthy: Exclamation marks are for exclamations.
235.
by glial 3y ago
Until the concentration is high enough it affects birth rates.
236.
by glial 3y ago
The definition of an 'intellectual task' used to mean 'abstract from experience' (Aristotle) or 'do symbolic processing' (Leibniz). Computers can now do these things - they can integrate better than Feynman, di
237.
by glial 3y ago
Sure, but if you continue: > they showed in live, behaving animals that they could enhance the response of visual cortex neurons to visual stimulation by directly controlling the activity of astrocytes. Perhaps we're talking past ea
238.
by glial 3y ago
The role of astrocytes in neural computation is an example. For a long time, the assumption was that astrocytes were just "maintenance" or structural cells (the name "glia" comes from "glue"). Thus, they were n
239.
by glial 3y ago
> The human brain is the only thing we can conclusively say does run a general intelligence That's because we define "general intelligence" circularly as "something the human brain does."
240.
by glial 3y ago
> We don't need to simulate them. You might be right, but this is the kind of hubris that is often embarrassing in hindsight. Like when Aristotle thought the brain was a radiator.

More ›