Lerc

42 ms·

841.
by Lerc 11mo ago
The same reason you need transistors to make computers. You don't need them, but they are one way to do it that people know how to implement. Identifying patterns is fairly amenable to analytic approaches, interpreting them, less so.
842.
by Lerc 11mo ago
In that respect I suspect it is much the same as bored apes. The price can go up while there are people with funds to put into things they don't care about. When the time comes that they have less money than the cost of things import
843.
by Lerc 11mo ago
It is, but only for thing people would legitimately like to have. The entire NFT thing would work if it were restricted to things people want, even if that only amounts to bragging rights. Somewhere along the way people lost track of the fa
844.
by Lerc 11mo ago
There are separate tests for the extremes, but obviously less researched because the further out you go the less they have to work on. Many years ago, while unemployed, I was sent for a intelligence and dyslexia test (because of the very sa
845.
by Lerc 11mo ago
As a New Zealander, I have often wondered about the standard deviation of sheep intelligence.
846.
by Lerc 11mo ago
Yes, but focusing on that is pretty much the opposite of the intent of the article.
847.
by Lerc 11mo ago
Quite. I have yet to manage a verification between clients. I have had all variations of clients ignoring requests, reporting requests only for the requesting client to ignore the response. Both ends quitting declaring that the other end
848.
by Lerc 11mo ago
Algortithms: State space models, diffusion models, KANs, hierarchical attention. There are no shortage of ideas. Determining what works well is a process that is going on right now. The question on planning on HBM is too vague to really a
849.
by Lerc 11mo ago
I think this is likely to happen in all models since their internal reasoning is not in the same form as the output. This is probably true also for humans. This may solve the additional clouding that comes from LLMs using what is an effec
850.
by Lerc 11mo ago
There seems to be a notion of 'Too big to fork' I don't think it is true. Certainly it takes more work. Broad consensus may be part of that work. If you cannot reach consensus to produce an equal product as a unilateral decis
851.
by Lerc 1y ago
Perhaps but a project using this stops you from writing any old C++ in your C. Writing C++ in a C style has no such protection. It's choosing which features are allowed in.
852.
by Lerc 1y ago
I feel like it did, it was always a weird combination of hacky things and entrepreneur things, and people writing to learn how to write. I find the tone of the posts also changes by time of day. I often wake up in the middle of the night
853.
by Lerc 1y ago
What counts as data center water consumption here? There are many ways to arguably come up with a number. Does it count water use for cooling only, or does it include use for the infrastructure that keeps it running (power generation, maint
854.
by Lerc 1y ago
I think the discussion on what standard practice should be does need to be had. This seems to be throwing blame at people following the current standard. If the obscure coded is not included by default or cannot be triggered by any means ot
855.
by Lerc 1y ago
That is standard practice. It is considered irresponsible to not publicly disclose any vulnerability. The X days is a concession to the developers that the public disclosure will be delayed to give them an opportunity to address the issue.
856.
by Lerc 1y ago
Given the focus on parallelism and communication, maybe the Connection Machine 2?
857.
by Lerc 1y ago
It's something to remember when considering the reactions and attitudes of extremely high profile people. If you've ever felt bad because of a shitty comment posted by someone you don't know, you have just a fraction of wha
858.
by Lerc 1y ago
My namesake was(died a few years back) the wealthiest person in the city I live in. He was worth peanuts compared to my brother's namesake who made billions through some sort of Lisp derivative shenanigans(or something along those lin
859.
by Lerc 1y ago
Which says. Update (August 13, 2025, 03:57 GMT): While the community correctly identified a performance issue, their attribution of the cause was mistaken.
860.
by Lerc 1y ago
Is there any significant difference between that and element.innerHTML=` <p>Today is ${new Date().toDateString()}</p> <p>Random number: ${Math.floor(Math.random() * 100)}</p> ` (ide
861.
by Lerc 1y ago
This should not be considered an LLM issue. It's a user experience issue. Mozilla has made changes that happened by default before. Often I have had to find the setting to put it back to how I wanted. I remember when it moved the URL
862.
by Lerc 1y ago
I don't think this shows that you can't trust things. I think it means trust should be earned. We might be transitioning to a world where trust has value and is earned and stored in your reputation. Clickbait is a symptom of pe
863.
by Lerc 1y ago
In that case I think the best advice I can give here is to focus less on features you dislike in other things and conaider the problems caused by those things. Without being encumbered by legacy reqirements you are free to make any changes
864.
by Lerc 1y ago
This is what I meant. I also like the idea of optical, line-of-sight connections. If you do the hypercube topology everything a node connects to has a different parity, so you lay them out on two panels facing each other.
865.
by Lerc 1y ago
I'm actually ok with that if it truly serving the purpose for what a computer should be. I think those principles would embody the notion that the same thing cannot serve all people equally. Simultaneously, for people to interact, int
866.
by Lerc 1y ago
I'm interested in the idea of a clean slate hardware/software system. I think being constrained to support existing hardware or software reduces opportunities for innovation on the other. I don't see that in this project. Thi
867.
by Lerc 1y ago
Perfect is the enemy of good, again. Many of the my best written projects never got finished. Some of my finished projects were hideous to behold on the inside. I did once do something that shipped that I was particularly proud of how it wo
868.
by Lerc 1y ago
I have aphantasia in the sense that I have no sense of there being an image, but several years ago, I posted a comment hypothesising that the interpretation of the experience as an image might be the distinguishing factor. The response to t
869.
by Lerc 1y ago
I feel like there are so many factors here that it's hard to identify which thing have the greatest impact. Instead of attempting a coherent argument, I'll offer some further observations. Taylor Swift is one of the most famous p
870.
by Lerc 1y ago
That requires tool use or some similar specific action at inference time. The technique I suggested would, I think, work on existing model inference methods. The ability already exists in the architecture. It's just a training adjustm

More ›