Scene_Cast2

6 ms·

1.
by Scene_Cast2 6d ago
How many tokens are you getting, roughly?
2.
by Scene_Cast2 12d ago
I found LLMs circa 4 months ago to be decent at textbook-style electronics, but they choke at anything nontrivial. They approach things like someone that knows the undergrad and grad textbooks, but has absolutely zero experience. No ability
3.
by Scene_Cast2 15d ago
Why does it work for Red Bull?
4.
by Scene_Cast2 16d ago
It worsens the supply crunch, no? A unit you use sparingly vs that memory going into a GPU that serves many more people.
5.
by Scene_Cast2 27d ago
If you need that precision, especially for position and not values, why use floats at all?
6.
by Scene_Cast2 29d ago
Explicitly trying to re-identify data that has been de-identified is typically a fireable offence at FAANG. Accidentally making a machine learning system that happens to (potentially) do it is a different matter.
7.
by Scene_Cast2 1mo ago
Oh hey, that's cheaper than Kimi K3! Amusing to see a SOTA OpenAI model be cheaper than a Chinese open weight model. Fwiw I love K3 and use it as a daily driver. I haven't tried Sol, as I dislike OpenAI.
8.
by Scene_Cast2 1mo ago
I made the move some time ago. There are some unpolished parts here and there, but I'm glad I made the move. Somewhat unrelated criticism of Gmail - during the migration, I found out that there is no way to search for starred archived
9.
by Scene_Cast2 1mo ago
Awesome. I'd love if Flutter could also finally get Impeller on Web (I'm tracking the GitHub issues, seems like there's progress at least). I personally would like for it to pick up some ideas from Svelte.
10.
by Scene_Cast2 1mo ago
In one of their dev surveys, they did float the idea of divesting Flutter.
11.
by Scene_Cast2 1mo ago
There's a trade-off between how long you wait and the skill spread of the team. IIRC at the very top, for DOTA2, the pros complained enough for the devs to make a "wait longer" matchmaking mode.
12.
by Scene_Cast2 1mo ago
I got a similar looking Stax amp (the SRM series). Also quite satisfying. It has a split volume dial so that you can adjust the volume of each channel separately.
13.
by Scene_Cast2 2mo ago
I wonder how much effect the CPU's register remapping has on these results. Also, they ran it on an ancient CPU, but I'm not sure how much that matters.
14.
by Scene_Cast2 2mo ago
I love the hackability of Thinkpads circa that generation.
15.
by Scene_Cast2 2mo ago
I'm using Kimi K3 + OpenCode. I pay their API pricing, costs about $5 / hour (and chews through ~10 million tokens / hour) during continuous use when I have one or two sessions running and doing their thing. Can't commen
16.
by Scene_Cast2 2mo ago
Very much agreed. Even more basic than that - memory access patterns are important. The amusing thing is that you end up writing GPU-style code even for CPU. For example - instead of an array of objects, using parquet-style object of arrays
17.
by Scene_Cast2 2mo ago
What are some good tools for _authoring_ HDR photography? Lightroom still seems to be mostly SDR. The new Resolve will probably have the tooling, but I haven't tried it.
18.
by Scene_Cast2 2mo ago
I meant that they're increasingly converging to be Oppo phones (now running the same OS, hardware is a slightly tweaked Oppo phone variant, etc).
19.
by Scene_Cast2 2mo ago
They were one of the brands with unlockable bootloaders and slide switches for mute. Unfortunately the Oppo takeover didn't preserve either. Written on a OnePlus 8 Pro.
20.
by Scene_Cast2 2mo ago
I wonder if this tool can help with EMC compliance testing. My TinySA needs an LNA, so I wonder if this has the required noise floor.
21.
by Scene_Cast2 2mo ago
The FIRE net worth seems unrealistically low.
22.
by Scene_Cast2 3mo ago
Huh. When I rented a Z7 ii about 5 years ago, I found their Android app to be pretty great. (My next big camera is likely to be a Nikon, in part due to the nice app)
23.
by Scene_Cast2 3mo ago
I've spent the last several years of building a very, very fancy scale (well, more of an industrial and R&D lab tool that can be used as a scale). The reason is that most people use a dirt cheap HX711 or cheaper. A fancier microcon
24.
by Scene_Cast2 3mo ago
I remember a story on HN from a while back. The idea is that the larger the org, the simpler the message and the tool has to be to reach everyone. The comment author was saying that as a junior, his company implemented a "tokenmaxxing&
25.
by Scene_Cast2 3mo ago
Not necessarily. I'm a proponent of (admittedly not very popular) methodology of "train, do interpretability analysis, adjust model architecture". It's not more popular for a few reasons: 1) you first need to train a ful
26.
by Scene_Cast2 3mo ago
IIRC the original author of the Lottery Ticket Hypothesis now disavows that idea. One intuitive way of looking at it is like so - let's say that you have a gaussian-looking plot. You want to fit a gaussian. You have a stupid simple mod
27.
by Scene_Cast2 3mo ago
Yes it would. Or, rather, labeling (not extra tokens).
28.
by Scene_Cast2 3mo ago
Really neat findings. I've personally had a line of thought where you bake in the role into the token. Basically have an embedding (same dim as token dim) for each role, add it to each token. This adds an unambiguous, unspoofable tag.
29.
by Scene_Cast2 3mo ago
I think this also stems from ML being more like biology or alchemy and less like math or programming (where you can get down to the first principles, abstractions are rock solid, and non-determinism is limited in scope).
30.
by Scene_Cast2 3mo ago
Krazam already has a video covering this exact idea.

More ›