leobg

13 ms·

151.
by leobg 1y ago
Exactly. > What he does is evil. Why is it evil? Because I resent him. Why do you resent him? Because I am no match for him. - Has anyone ever answered like this? (Nietzsche)
152.
by leobg 1y ago
A zero effort, dismissive, libelous comment, driven by an emotional reflex befitting the poster’s name.
153.
by leobg 1y ago
Brave thing to say on HN. There are a few people here who will downvote any comment that contains the word “Elon”, “SpaceX” or “Tesla” if the comment’s saltiness score is less than 8/10.
154.
by leobg 1y ago
> Often someone pronounces his propositions with such confident and inflexible defiance that he seems to have entirely laid aside all concern for error. A bet disconcerts him. Sometimes he reveals that he is persuaded enough for one duca
155.
by leobg 1y ago
Many ppl think to fine tune an LLM on domain knowledge means to feed it chunked text of, say, psychology books. That is, of course, a wrong application if your goal is for the model to become an expert psychologist. You want the behavior of
156.
by leobg 1y ago
Well said. And it makes sense, if you define intelligence as the ability to successfully predict the future. And how interesting that that is literally how LLMs are trained during pretraining. Like Ilya said: To predict the name revealed as
157.
by leobg 1y ago
Free, dost thou call thyself? Thy ruling thought would I hear of, and not that thou hast escaped from a yoke. Art thou one ENTITLED to escape from a yoke? Many a one hath cast away his final worth when he hath cast away his servitude. Free
158.
by leobg 1y ago
Might also be useful for dataset curation, or even just prompt engineering. For example when training a classification task and picking a diverse set of examples for training or evaluation.
159.
by leobg 1y ago
Is it just income that’s the issue? I’d rather say it’s purpose. Even more: What will happen to democracy in a world where 100% of the population are 27/7 consumers?
160.
by leobg 1y ago
I’ve read that same sentiment back in 2016. Only difference being that “meme stock” wasn’t a word back then. I also love the excuse: “I’m right. I’m just not putting any money behind my conviction because the market might deviate from my tr
161.
by leobg 1y ago
For the impatient (and those without ad block): > SPR AVS (Spatial Relay Audio-Visual Sync): Apple's new wireless protocol that enables lossless audio transmission capability—something Bluetooth fundamentally cannot do. Uses the N1
162.
by leobg 1y ago
A fellow Nietzschean? :)
163.
by leobg 1y ago
Is there a binding precedent (like a BGH or OLG ruling), or is that just someone’s theory that it’s not covered? I did not check. If there is no binding ruling, your Amtsgericht judge may see things differently. Even if the law didn’t cover
164.
by leobg 1y ago
I was about to buy a NAS. I find the idea of using an old laptop instead interesting. Especially since it comes with UPS built in. The author is using a ThinkPad T430. Any experiences?
165.
by leobg 1y ago
I’d think this is illegal since 2022. You may want to check. https://www.bmjv.de/SharedDocs/Pressemitteilungen/DE/2022/02...
166.
by leobg 1y ago
What have you done in your own home? I’d be curious to know.
167.
by leobg 1y ago
Seems like you and Adam Smith put different weights on the value of altruism. > It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.
168.
by leobg 1y ago
Beautiful. Sounds like they’re doing with pens in Maryville what Tesla did with the Model S in Fremont. > Peterson […] found that the factory could use robots to do an increasing share of the packing. But he decided to keep the employees
169.
by leobg 1y ago
Of course. Because the unseen part here is that the model is being taught that every other representation of the same fact was wrong. Meaning, during training, if the model expresses the same fact in some other form, maybe even with just on
170.
by leobg 1y ago
What model is that? Ink doesn’t dry and lasts for five years when you’re going through 500 pages a month. Sounds incredible.
171.
by leobg 1y ago
Retrieval Augmented Generation. How you do the retrieving is irrelevant. You can do it manually and it’ll still be RAG. Also, most RAG pipelines combine multiple approaches - BM25, embeddings, etc..
172.
by leobg 1y ago
litellm + gemini API key? I find Gemini is their first API that works like that. Not like their pre-Gemini vision, speech recognition, sheets etc.. Those were/are a nightmare to set up indeed.
173.
by leobg 1y ago
Not an exact match, but I liked this one a lot (read it twice): "Faster, Better, Cheaper" in the History of Manufacturing: From the Stone Age to Lean Manufacturing and Beyond by Christoph Roser
174.
by leobg 1y ago
So basically kokoro, but VC backed? Both models use espeak, so it seems like it’s the same general approach. Demo sounds great.
175.
by leobg 1y ago
It’s a theory. But what if it’s wrong? > Second is by clear communication by European leaders, that any act of sabotage, aggression and other acts of hybrid warfare will have analogous and larger scale responses on the Russian soil. You
176.
by leobg 1y ago
Direct link: https://ycarena.com/
177.
by leobg 1y ago
No, not really. I didn't see the cultural reference. Never heard of Baudrillard before. I was just playing off the word "revenge". That being said, your take did stop me in my tracks. Even though it's not the first time
178.
by leobg 1y ago
This reads like an ad for Arbitrus.ai. It’s copywriting lingo: > We built one called Arbitrus. We put it through a mini-Choi test and it mopped the floor with the competition
179.
by leobg 1y ago
German: “Klammeraffe” (literally: “bracket monkey”)
180.
by leobg 1y ago
Would have been ironic if it was the crawler from OpenAI… :)

More ›