pantelisk

4 ms·

1.
by pantelisk 10d ago
I think the doom demo uses a text representation of the world and it's basically, "projectile coming your way" -> "Strafe". "Enemy ahead" -> "shoot. So it works well when spawned in a room of en
2.
by pantelisk 12d ago
But how did they know about it? Did the agents independently discover it? Did the model know about it because this vuln was public somewhere and systems weren't patched yet?
3.
by pantelisk 23d ago
Same for smartphones. There was a time of unlimited hype and secrery around new iphones. Who remembers the story of an iphone 5 prototype left a bar. Journalists were going crazy, people were signing petitions for Apple to not hunt down but
4.
by pantelisk 23d ago
Audacity :) (as an app) and Audiomass on web (I am biased towards it of course)
5.
by pantelisk 1mo ago
It should look like a NES cartridge! That you have to blow on its end to clear any dust and it should do a satisfying click when it slots in. Cooling might be an issue though...
6.
by pantelisk 1mo ago
This is part of the reason I wanted to announce it even before I feel ready to ship... The frontier models have such crazy velocity (admirable) and moving so fast that I didn't want them to make an announcement and suddenly trivialize
7.
by pantelisk 1mo ago
So, this is what I have been thinking (and part of the reason why I want to "cook" it more before shipping). Since I have been building it for so long, I am not sure how much "overfit" it is on my data, the way I speak a
8.
by pantelisk 1mo ago
Yes, the dog has a few levers including a "boredom" counter that will trigger it to do things on its own (like roam around, eat something, chase something (eg a rabbit) if there's one around etc). All connected with its curre
9.
by pantelisk 1mo ago
At that point it's deterministic and doesn't need the LLM. Once a plan has been decomposed and evaluated, it becomes a sequence of actions. Actions can either wait on a trigger or flow from one to another. For the arrow example: w
10.
by pantelisk 1mo ago
I will be open sourcing it soon :) my code is a bit dirty (the whole system is 3 pieces. The game adapter / a websocket bridge between game and brain / the brain itself) and it runs on 2 different (local) machines currently. Ale i
11.
by pantelisk 1mo ago
Hello, author here. I was intentionally a little vague about this because this is kind of the thing that makes the whole approach work. An LLM predicts the next token. If you're trying to predict the next token in a mathematics competi
12.
I built a low-latency AI companion that plays Skyrim with me(pantel.is)
399 points by pantelisk 1mo ago|76 comments
13.
by pantelisk 2mo ago
While I agree with the premise, I think this angle only applies on work one was going to do no matter what. The real power of these tools is that there are so many ideas people would like to try, but never have the time or motivation to pur
14.
by pantelisk 3mo ago
Really fun project, but doesn't take ageism into account, it gets easier the further you get whereas it should get be getting harder in certain ways
15.
by pantelisk 3mo ago
There are also solutions for sharing your homelab with others (basically tunneling from your machine->server (internet accessible) <-> client. Though, if your machine would go to sleep that whole chain would fall apart. A few good
16.
by pantelisk 3mo ago
Yes, you can see an early demo video in the page, of a very early version (I decided not to post the final wip yet). I’m building an AI gaming companion. Essentially an AI NPC (a speaking dog, think dogmeat of fallout on steroids) that is g
17.
by pantelisk 3mo ago
Hi HN — I’ve been working on a real-time conversational AI, and one of the biggest problems is speech latency. As I was experimenting with various TTS systems I stumbled upon Pocket-TTS by Kyutai. It runs on the CPU which made it a great ca
18.
Show HN: PocketTTS-raven – fast local in-browser TTS with voice cloning(pantel.is)
4 points by pantelisk 3mo ago|4 comments
19.
by pantelisk 3mo ago
And I think both have seen these much older handcrafted css3d engines - https://pantel.is/projects/css3d/ - https://keithclark.co.uk/labs/css-fps/ (the original) but quake and doom took
20.
by pantelisk 4mo ago
OpenDAW is different beast in the same field. It's a full music editor and suite, where you can write notes down on a piano roll (or record with midi), then use synthesizers and samples to turn that into a song. Audiomass is an audio e
21.
by pantelisk 4mo ago
oh to go back to multitrack, there is a button at the top right (multitrack beta). Maybe it feels abit disconnected being there
22.
by pantelisk 4mo ago
I just tried it with a 2 hour podcast file, since more people asked about long files. It is performs great on chrome on my pretty powerful macbook. Some operations are a little slow and zooming in to max level starts to be slow, default thi
23.
by pantelisk 4mo ago
Thanks for your comment :) you can press "preview" (and there is a little on/off) in the effect's modal window. But I agree with you, an automation type system that operates on the entire track might be better
24.
by pantelisk 4mo ago
Author here. License is pretty much do whatever you want with it (free as in free information and free beer), I suppose the closest one to that would be MIT, but I don't like its serious legalese tone. I prefer the whimsical "free
25.
by pantelisk 4mo ago
you can click and drag to select part of the audio (and then drag the edges of the selection region that has appeared if you want to adjust it), and then apply the effect. All effects prioritize the current selection first, and if no select
26.
by pantelisk 4mo ago
Hey HN. Thanks again for featuring this project :) this is my favorite community on the internet and my go to site to visit, almost by reflex, when I have idle time, so it feels me with joy to have made something well received here. I &#x27
27.
by pantelisk 4mo ago
Oh.. I 'm mindblown! I was wrong, this is the great tracker we were missing
28.
by pantelisk 4mo ago
I think this git but for music you are suggesting is quite interesting (or more like a figma for music maybe). My musician friends still use dropbox and google drive to push around files to each other. Honestly, I would be all for it but I
29.
by pantelisk 4mo ago
Oh, that's fantastic to know. Did you load it into multitrack, or the regular editor? (in multitrack it does not scale currently, but working on it). On regular editor it should in theory try to zoom. There is a pyramid cache mechanism
30.
by pantelisk 4mo ago
All very good points, not much to say I agree with you. With loading plugins on demand it could grow in size without affecting load and experience (and since offline mode is a separate link that would still be fine to be a little larger sin

More ›