JasonSage

6 ms·

1.
by JasonSage 27d ago
There's an exceptional idea buried in this which flies under the radar in my experience: If we're writing a document that polices the model behavior, we're altering every version of the model+harness we use henceforth. Most t
2.
by JasonSage 2mo ago
Yeah, I mostly mean for the TTS, which in my testing is flat and has no emotional register, or no nuanced variation. OpenAI speech in my experience is very good at things like communicating understanding or questions non-verbally using into
3.
by JasonSage 2mo ago
I for one am greatly looking forward to the day these kind of voice models can be run locally. It seems like the gap between open-weight and frontier is way larger for voice models than coding/language models.
4.
by JasonSage 3mo ago
It's also a great way to not become the next "graphics programming legend" --I think a fast-moving field with lots of new developments is actually an exciting place to be a pro.
5.
by JasonSage 3mo ago
I think the situation with anti-cheat on Linux is changing. Studios are putting resources into anti-cheat that will work on Linux. If I'm being a bit cynical, I could say this is "just" because of Steam Deck and Steam Machine
6.
by JasonSage 3mo ago
I appreciate that a lot! The article has a deliberate and explicit scope, and covers it well. I'm hoping that perhaps my personal perspective shades why "instances" comes up, or why the reaction on HN seems to include the wid
7.
by JasonSage 3mo ago
From my perspective, I care about the centralization/decentralization aspect a lot, and if I'm coming into the discussion with a much better understanding of the Mastodon side then _of course_ I'm going to ask about the insta
8.
by JasonSage 3mo ago
I keep a bookmarks folder of websites that have non-cookie-cutter design. If I have an aesthetic in mind I'll use some screenshots of those sites in the prompt and phrase their inclusion as: "Look at these slightly non-standard de
9.
by JasonSage 3mo ago
I've had that experience and I've started testing different prompts and inputs. I find it funny about meeting requirements when you give them, and making safe choices when you don't give direction. So if you're going to
10.
by JasonSage 3mo ago
> content co-authored with AI that is carefully reviewed The problem I see is that this is indistinguishable to a reader at a glance. Distancing the writing from the "AI smell" not only improves the quality by dropping the unne
11.
by JasonSage 3mo ago
When there's no discernable human filter on the text output, reading the text suggests it's what the LLM produced and not what a human considered. This is low-quality--every single day I witness Codex and Claude misunderstand, mis
12.
by JasonSage 5mo ago
I see it as having a blue component and a green component. If the mixture has more green than blue, then it's green. The analogous version in black and white is "is this dark grey or light grey?" because that's the one a
13.
by JasonSage 6mo ago
I enjoyed this, though my font preferences are pretty stable. It would be nice if it showed you 1st, 2nd, semi-finalist, quarter-finalist... It would also be nice to see progress of some kind, a few minutes in I was wondering if I was near
14.
by JasonSage 7mo ago
THINKalpha | Senior Frontend Engineer | Remote | Full Time We're building alphaLENS ( https://alphalens.io ), a suite of tools for professional market research and trading. Public registration opens in days. You can see our c
15.
by JasonSage 8mo ago
Do you think nobody felt that way about cars?
16.
by JasonSage 8mo ago
Yes, we're getting errors in the checkout action, and I haven't been able to git fetch for 5-10m.
17.
by JasonSage 8mo ago
ChatGPT recommended this to me recently when I was trying to get some assistance with a usable Tailwind palette. I ended up not needing it right away but it's first in line next time I need to make one.
18.
by JasonSage 9mo ago
You're right. People want to pretend fundamentals of economics don't exist AND the company has moral obligations to fulfill to consumers. It's laughable. It's not just nVidia, I've seen other expensive consumer bran
19.
by JasonSage 9mo ago
I suspect that the non-Rust improvements are vastly more important than you’re giving credit for. I think the go version would be 5x or 8x compared to the 10x, maybe closer. It’s not that the Rust parts are insignificant but the algorithmic
20.
by JasonSage 9mo ago
Yeah, that’s a huge over-generalization
21.
by JasonSage 9mo ago
Local non-utility does not imply global non-value. Of course there's costs and benefits, but it's hard to have a conversation with good-faith comparison using "many see it as overly complex" -- this is an analysis that c
22.
by JasonSage 9mo ago
> Composed resolvers are the headache for most and not seen as a net benefit, you can have proxied (federated) subsets of routes in REST, that ain't hard at all Right, so if you take away the resolver composition (this is graph comp
23.
by JasonSage 9mo ago
I'm building a yet another AI chat app. My initial goal is to make a functional SillyTavern (AI roleplaying) replacement. SillyTavern builds prompts from a few rigid buckets (character, scenario, lore, system prompt, author's note
24.
by JasonSage 10mo ago
It sounds like you want to make a judgement call: "they're too small to enforce this license agreement," so you get to pretend it's an honor system and not a license agreement.
25.
by JasonSage 10mo ago
When there's a license you're either violating the license agreement or you're not. That's not an honor system.
26.
by JasonSage 11mo ago
My question is what happened between when they went in the water and when they got off-site medical treatment. 7 hours seems like a long time. Is there on-site medical that would be doing something during that time?
27.
by JasonSage 11mo ago
But "it's not possible for xxx to be used securely" is a better premise if it deflects people who can't do it correctly.
28.
by JasonSage 11mo ago
That’s fine.
29.
by JasonSage 11mo ago
I have to agree. My experience working on a team with mixed levels of seniority and coding experience is that everybody got some increase in productivity and some increase in quality. The ones who spend more time developing their agentic co
30.
by JasonSage 1y ago
To me the biggest upside would be terminal completion and discovery via help text. Sure you can always bounce to a search engine and bounce back, but I can imagine cases where you want a toolkit in front of you that you know how to use when

More ›