Aurornis

39 ms·

871.
by Aurornis 4mo ago
It is preferable if genetically engineered crops cannot reproduce. The impact of their wild propagation should be studied first, not unleashed upon the local ecosystem.
872.
by Aurornis 4mo ago
This argument has been beaten to death before AI: Ever since calculators were able to do math, students have been wondering why they need to learn how to do all of this math manually when they could get the same answer from a calculator. Th
873.
by Aurornis 4mo ago
SSD streaming throughput is too slow to be usable. GLM-5.2 has 40B active parameters at a time. At Q4 that's 20GB. The best PCIe 5 SSDs can get 15GB/sec when everything goes well. Every expert load would take more than a second. I
874.
by Aurornis 4mo ago
I think you missed the part where this is a lossy technique that reduces performance. The image trick reduces context because it’s lossy. The README says you can’t use it for anything needing exact recall. It produces a gist of the input. Y
875.
by Aurornis 4mo ago
> hedge against the various tail risks of third-party providers raising prices They could 10X the prices and you’d still be better off. It’s also unlikely that prices go up enough to warrant a $100K local investment to prevent paying a c
876.
by Aurornis 4mo ago
> It'll almost certainly be worth it, given the abusive behavior we've seen and will continue to see from the major closed-model providers. The proper financial comparison for GLM-5.2 would be one of the providers on OpenRouter
877.
by Aurornis 4mo ago
I have an M5 MacBook Pro and I also have a separate GPU setup for running models. The difference in speed is significant. It's not just token generation speed, but time to first token (prompt processing). The M5 hardware is amazing for
878.
by Aurornis 4mo ago
I play with local LLMs a lot. I've spent more on hardware than I should. I'm friends with a local group of people who have spent a lot more than I have. The warning I would have for everyone is to temper your expectations and read
879.
by Aurornis 4mo ago
> They universally tell me that their coworkers generate awful PRs with bugs, nonsensical code and fake unit tests. But they also universally tell me that they are different, have special workflows and prompts that create good code If so
880.
by Aurornis 4mo ago
> State founded school is not going to tell you it costs $1200000 to raise baby Completely wrong. Not even plausible. You think every family is spending $67K per year, per baby, until they're 18? What?
881.
by Aurornis 4mo ago
> I have no idea if Switzerland is any better, but the US situation in 2026 is appalling. Kind of amazing that we're calling 1Gbps fiber "appalling". Every thread about internet access attracts people with unique situation
882.
by Aurornis 4mo ago
This article came up before. It's heavy on the clickbait, if you couldn't guess from the title. Some important points that they leave out: - 25G internet isn't available everywhere in Switzerland. It's just the fastest
883.
by Aurornis 4mo ago
You're not missing anything. The article is trying to imply that 25G is everywhere in Switzerland, but it's not. They just picked the fastest available speed and ignored the fact that some places in the United States also have 25G
884.
by Aurornis 4mo ago
> Specifically: A pipeline from the ocean. The environmental cost of building a pipeline 750 miles across the country and then expending all of the energy needed to pump that water would completely outweigh any benefits. You also don&#x2
885.
by Aurornis 4mo ago
The avian flu pandemic started in early 2022. I looked up old news articles and a lot of them are dated February 2022. It really did increase egg prices. The collusion quotes in this article start in October 2022, though it doesn't say
886.
by Aurornis 4mo ago
> I went through two iterations of reaching out to my college network. The first time I put so much time into handwriting notes and trying to provide my relatable background. 100 notes, not a single response. This is a thoughtful gestur
887.
by Aurornis 4mo ago
> One of the strongest ways to show that you’re worth helping is to demonstrate that you are a serious person When I’ve been in positions where a lot of people ask for help, this is the #1 place I saw people drop the ball. The advice to
888.
by Aurornis 4mo ago
> As everyone should know by now, it is not in general possible to find bugs by examining the code. Lost me here. I would agree that it’s not possible to find every bug by examining code, but in real code reviews bugs and errors are id
889.
by Aurornis 4mo ago
> Our entire small team thumbs up a PR before it's merged unless there's a big rush on it This has been tried by a couple of my past managers. This feels great with a small team on a slow moving codebase. If you try to force it
890.
by Aurornis 4mo ago
Which requires some record keeper transaction. Which turns into a handshake with the centralized entity, the government. If every token is single use then you also need to get them all from the government, either on demand or in bulk. They
891.
by Aurornis 4mo ago
A past employer thought it was a good idea to put up a leaderboard of who sent the most Slack messages. They celebrated the people at the top for being so active. Predictably, everyone started talking in Slack like their jobs depended on it
892.
by Aurornis 4mo ago
> But if they reveal nothing, isn't it wide open for abuse? Couldn't one over-18-person's proof become everyone's proof, because they can't tell it's the same proof, and the issuer can't tell where or h
893.
by Aurornis 4mo ago
> I know that with oscilloscopes, it’s recommended to use 5x instead of nuquist 2x of the highest frequency you want to use. For capturing analog signals, 2.5X is enough headroom. The 5X recommendation is probably for digital signals whe
894.
by Aurornis 4mo ago
Rust is worth learning, but your C skills will continue to be useful for embedded for a long time. Rust support in the embedded world is still growing and you will find yourself going between Rust and C on most projects unless you can caref
895.
by Aurornis 4mo ago
We’re talking about SOTA models like Fable, though. If you’ve got a product where the budget allows for Fable level token costs, I doubt you wouldn’t have the budget to run your evals again on a cheaper model if Fable was unavailable. I mea
896.
by Aurornis 4mo ago
I have the luxury of being relatively close to my kids’ schools and a grocery store and I’m also fit enough to bike and walk long distances. The only reason I can do this is because I work remote and we have the means to purchase a house al
897.
by Aurornis 4mo ago
> We were on forums (even usenet) and got along. Now it is all walled gardens, rage bait, racism, and people shouting at each other. You’re remembering the good parts and forgetting the bad parts that you looked past at the time. Old use
898.
by Aurornis 4mo ago
> I think everything changed when Eternal September happened. The original Eternal September is about a specific year, but it has become an evergreen concept for each younger generation: The Internet was really cool when I started usin
899.
by Aurornis 4mo ago
SOTA models are typically used for interactive coding and other human in the loop work > say GPT-4o to GPT-5.2, a transition I just finished on a not too complicated application Neither of which is close to SOTA, because tasks like these
900.
by Aurornis 4mo ago
> The damage is done. You cannot build a business critical function on top of American SOTA frontier model. Especially not with the current crew in charge. The switching costs of changing LLM providers is as low as it gets. All the indiv

More ›