Viewing profile — hansihe

They are not really token-in token-out per se, they are embedding-in embedding-out. When operating on text, you embed each token into the LLMs embedding space. You go from a discre…

It seems likely to me this was driven by the `ultra` mode in 5.6, which fans subagents to do work. This mode was previously only available in the web UI (what was previously known …

What do you mean? There is nothing unenforceable about this.

CockroachDB does Serializable by default

What makes it so great for me is the effortlessness. I often use Python for quick one off scripts. With UV I can just do `uv init`, `uv add` to add dependencies, and `uv run` whate…

Your dismissal of this research shows a fundamental misunderstanding of how scientific analysis works. The study analyzed two separate phenomena - changes in engagement for Musk's …

The study analyzed two separate things: * Change in engagement for posts by Elon Musk * Change in engagement for republican content It seems to me that when taken together, these t…

Why is that laughable exactly? It was the same point in time which he endorsed Trump

It's worth noting that when Trump says "radical left lunatics" he means mainstream democrats like Kamala Harris. https://www.aljazeera.com/news/2024/7/25/trump-blasts-harris...

I think it aged perfectly well. We were talking about his son, he let the process play out in full which is beyond respectable. When the incoming administration has shown extraordi…

After the precedent around pardons set by his predecessor I'll give him that one. At least Biden let the process play out before issuing the pardon so the public got to know all th…

$375 per box doesn't seem bad to me at all when you probably only need a couple per school?

It's still open source and actively maintained by Apple, they use it internally. https://github.com/apple/foundationdb

Edge TPUs are definitely not comparable to the datacenter TPUs. They only support TFLite for one.

I don't know about this at any detailed level, but doesn't designing standard cells for leading edge nodes involve a lot of trial and error? Is a lot of the issues that can occur e…

It's probably more of a node thing than a fab thing. You would have a much easier time getting the fab to do random stuff for you on a legacy node compared to a leading edge node. …

This is really cool! Do you have some more example images somewhere?

For cases like this, you just need to convince it that it would be inappropriate to generate anything that does not follow your instructions. Mention how you are planning to use it…

> We are writing to inform you that we have discovered two Home Assistant integration plug-ins developed by you [...] that are in violation of our terms of service > Specifically, …

I was confused by this article in the beginning, it does a pretty bad job at drawing a distinction between the pure quadlet example at the start and the example of using CoreOS to …

While that's true, you really only need to blast out RGB data to get an image on screen. Most of what you are talking about is layered on top and optional. I did a tiny HDMI implem…

Kagi has the ability to customize site weights which I really like.

Anecdotally this actually works. An LLM is trained on a wide distribution of data. "Just asking" it to not make mistakes makes it more likely to sample the part of the distribution…

That's not true, ChatGPT is a model. Quote from the ChatGPT announcement post: > ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in a prompt an…