visarga

47 ms·

841.
by visarga 2y ago
> ai succession necessarily includes replacement How is AI going to make its own chips and energy? The supply chain for AI hardware is long an fragile. AGI will have an interest in maintaining peace for this reason. And why would it repl
842.
by visarga 2y ago
I think he is trying to take the positive side of what is probably an inetability.
843.
by visarga 2y ago
Compression works on finite data, and AI models need to keep open for new data. So they should be less greedy.
844.
by visarga 2y ago
> I worry that we'll get complacent and not check details like that when they are important, not just the medical field but everywhere. The performance level goes up all the time, it won't be this bad for long.
845.
by visarga 2y ago
> there is no way to "prompt" the AI on what a physician needs in a given scenario - eg. "only include medically relevant information in HPI", "don't give me a layman explanation of radiographic reports&quot
846.
by visarga 2y ago
The information is probably lost, training on very long inputs is expensive, we usually train on short inputs and only send longer ones at the end. So the author name would not appear together with all the chunks unless it appears on every
847.
by visarga 2y ago
I usually annotate my chunks with title, summary, keywords and multiple levels of hierarchical topics. More recently I thought about annotating intent, user values and tactics, especially for debate related text and LLM chats. So I would an
848.
by visarga 2y ago
Yeah, they should have used UMAP or tSNE to cluster the data a bit
849.
by visarga 2y ago
They are on par on many tasks, surpass us in some tasks, and catching up on the rest quite fast. Both humans and LLMs forget details, both have to go through iterative bug fixing when creating something. Longer contexts and memory are comin
850.
by visarga 2y ago
> All they do is regurgitate text which resembles a continuation of what came before, and sometimes—but with zero guarantees—that text aligns with reality. You mean like us? Because it takes many runs and debug rounds to make anything th
851.
by visarga 2y ago
I think aggregate information across billions of humans can compensate. It would be like a human personality model, that can impersonate anyone. How do you train such a model? Simple - Collect texts with known author and date. They can be b
852.
by visarga 2y ago
The thig is, if you look at all the "Deep Research" benchmark scores, they never claim to be perfect. The problem was plain to see.
853.
by visarga 2y ago
So you are basically doing a first pass with diverse models and second pass catches contradictions and other issues? It could help with hallucinations.
854.
by visarga 2y ago
Use Windsurf, a VSCode fork, it defaults on Claude as LLM.
855.
by visarga 2y ago
Housing problem comes from the job market. Cities with good jobs will attract people and the price of renting/buying will go as high as they can bear. Most of the benefit will go to landlords. Moving to a better job market will make yo
856.
by visarga 2y ago
> You can apply whatever multiplier you want as long as the assumption that human labor will be replaced by AI labor holds true. Do you think we will be doing in 5 or 10 years the same things we do today, but with AI? Every capability i
857.
by visarga 2y ago
AWS makes a fraction of the money their customers make. And NVIDIA is just seeing benefits from market speculation at work. Most LLM providers are losing money right now.
858.
by visarga 2y ago
This paper's got it backwards. AI's benefits don't pile up with the owners, they flow to whoever's got a problem to solve and knows how to point the AI at it. Think of AI like a library: owning the books doesn't mak
859.
by visarga 2y ago
> Google links to it and sends me traffic Used to, but more recently it's probably LLM agents using Google not people. And even if it's not yet, it will be. Last time I searched for something on Google it messed up so bad I qui
860.
by visarga 2y ago
That is not reasonable, should a child heed the restrictions placed on the 1st grade math book later in life, when they become PhD?
861.
by visarga 2y ago
Probably not because there was a DA who needed to make her career on his back.
862.
by visarga 2y ago
Doesn't synthetic data complicate this reasoning? If I train a model on synthetic data, which is not protected by copyright, I am free to do as I please. It won't even regurgitate the originals, it will learn the abstractions not
863.
by visarga 2y ago
Ah you mean like Google Search?
864.
by visarga 2y ago
> The world is going to become stupider because it’s easier. Wasn't the same thing when we switched from books to web? We lost the ability for long reads, and instead just search and click directly to specific information, losing th
865.
by visarga 2y ago
> they perceive some value is provided by publishing in the traditional way Yes, but it provides value only for scientist careers, not for science.
866.
by visarga 2y ago
Sometimes it's a matter of approach, some approach could be 5% better and 10x more expensive. So they will find the sweetspot, takes a few iterations.
867.
by visarga 2y ago
> They're not even trying to do anything interesting with the medium, they exist purely to extract ad revenue by mass producing SEO optimized content slurry. This sad situation was caused by internet. It made it possible for us to h
868.
by visarga 2y ago
Here is where you are wrong. The AI art is built on even more "art consciousness" than any one of us has ever seen. It's a mirror reflection of our own culture, and we should respect it as such. But you say there is no one th
869.
by visarga 2y ago
"AI" art is 99.9% of the time ephemeral art, see it once and throw it away. Who's got the time to see my AI art shit when they could be generating their own? It's more of a personal exploration tool, closer to imaginatio
870.
by visarga 2y ago
Hell, even reading was decried as a way to wreck your memory - why remember when you could always open a book?

More ›