upghost

9 ms·

1.
by upghost 1mo ago
I love how first we destroyed the word "AI" now we want to destroy the word "telepathy"
2.
by upghost 2mo ago
I do indeed mis-using dashes.
3.
by upghost 2mo ago
I'm not sure why but Prolog seems to have more misinformation circulated about it than any other programming language I've ever seen. I'd take a lot of the information in these comments with a grain of salt. If anyone is inte
4.
by upghost 3mo ago
theprimagen called this[1] like three days ago. That was fast. [1] https://www.youtube.com/watch?v=m-bT5v5Tm7w&t=164s
5.
by upghost 4mo ago
If this wasn't CERN tech I would think I was being taken for a ride. Conventional wisdom is that distributed consensus is not possible at this kind of performance, does anyone have a sense for how this is different and how my mental mo
6.
by upghost 4mo ago
Thanks. I thought it was interesting choosing arithmetic instead of some other relation because multimodal arithmetic (via CLP) is more of a PhD thesis than a blog post. Other relations might've been easier to demonstrate a general que
7.
by upghost 6mo ago
Man that is such a bummer. The Naval Support Activity (NSA) "base" is not a hardened military facility. I've never been to the one in Bahrain, but it's usually where you go to play ultimate frisbee, maybe some paintball
8.
by upghost 6mo ago
Great writeup. Only thing I din't see in here was an analysis of the impact of players like Talaas[1] and their stupid faster hardware LLMs. I feel like it could be majorly disruptive, but idk if it's going to prolong the apocalyp
9.
by upghost 6mo ago
> familiarity vs simplicity Love this, I've never heard it put that way before.
10.
by upghost 6mo ago
Yeah, this checks out. I wonder what they are doing to prevent semantic collapse. Also, I wonder if the base model would already be instruct and RLHF tuned or only pre-trained. Trying to do additional training without semantic collapse in
11.
by upghost 6mo ago
> Pre-training allows organizations to build domain-aware models by learning from large internal datasets. > Post-training methods allow teams to refine model behavior for specific tasks and environments. How do you suppose this works
12.
by upghost 6mo ago
> We are doing this to self-fund further investment in AI and enterprise sales while strengthening our financial profile. Some quotes from the video: > ...at the same time, we're a people company. > Your work will live on in o
13.
by upghost 6mo ago
In Prolog the syntax is incredibly important. It is designed to be metainterpreted with the same ease in which a for-loop might be written in another language. https://www.metalevel.at/acomip/ mi1(true). mi1((A,B
14.
by upghost 6mo ago
Very interesting stuff. Apparently this is the implementation: https://github.com/dicpeynado/prolog-in-forth Thinking about the amount of thought and energy that went into this, back in 1987 -- mostly preinternet, pre-
15.
by upghost 7mo ago
I'm surprised how hard I had to dig for an actual example of syntax[1], so here you go. [1]: https://www.lix.polytechnique.fr/~dale/lProlog/proghol/extra...
16.
by upghost 7mo ago
So I have been doing formal specification with TLA+ using AI assistance and it has been very helpful AFTER I REALIZED that quite often it was proving things that were either trivial or irrelevant to the problem at hand (and not the problem
17.
by upghost 7mo ago
Anakin: I'm going to save the world with my AI vulnerability scanner, Padme. Padme: You're scanning for vulnerabilities so you can fix them, Anakin? Anakin: ... Padme: You're scanning for vulnerabilities so you can FIX THEM,
18.
by upghost 8mo ago
tl;dr - All this AI stuff is just Universal Paperclips[1] I see a lot of comments about folks being worried about going soft, getting brain rot, or losing the fun part of coding. As far as I'm concerned this is a bigger (albeit kinda f
19.
by upghost 8mo ago
Author: "Not my favorite language" Prolog: "Mistakes were made" As an avid Prolog fan, I would have to agree with a lot of Mr. Wayne's comments! There are some things about the language that are now part of the ISO
20.
by upghost 9mo ago
Ok what I would really love is something like this but for the damn terminal. No, I don't store credentials in plaintext, but when they get pulled into memory after being decrypted you really gotta watch $TERMINAL_AGENT or it WILL re
21.
by upghost 10mo ago
A really simple one is traversing a linked list (or any naturally recursive data structure, such as a dictionary or tree). It is very natural to traverse a recursive data structure recursively.
22.
by upghost 10mo ago
Really impressive! For anyone who's not a pythonista, trying to implement TCO is something akin to solving the Collatz conjecture but for Python. It's often just an exercise in madness. So seeing an elegant solution to this is rea
23.
by upghost 10mo ago
Ok hear me out. It's not particularly obvious to me why plants being easy to replicate suddenly destroyed the rare plant market. Surely they can't be easier to replicate than software. That hasn't seemed to put much of a dent
24.
by upghost 10mo ago
No, and its ok to be dismissive of it. I'm just giving an experience report. Actually my primary value is emacs integration with claude. I have an mcp with one function ( as_user_eval ) which allows it to execute arbitrary s-expression
25.
by upghost 10mo ago
Chrome Devtools is similarly an extremely high value MCP for me. I would agree that if you don't find they add discoverability then MCPs would have no value for you and be worse than cli tools. It sounds like we have had very opposite
26.
by upghost 10mo ago
So I don't disagree with any of the criticisms of MCPs but no one here has mentioned why they are useful, and I'm not sure that everyone is aware that MCP is actually just a wrapper over existing cli/API: 1. Claude Code is a
27.
by upghost 10mo ago
Taking the article's analogy of the "collaboration while driving", the F1 sport is quite insanely collaborative. For example, the drivers literally do have someone in their ear by radio being their coach and spotter for the e
28.
by upghost 11mo ago
Not the type theory we wanted, but the type theory we needed.
29.
by upghost 11mo ago
Comments here are a bit rough. Marcin does plenty of work for OSE and the community is still active but most of it is in-person/offline -- hard for a lot of us to believe, I know, but there is human activity that takes place outside th
30.
by upghost 1y ago
Ok to be fair the eagle looks just as confused and distressed to be there as the rest of us

More ›