ryeguy

They can benchmark it because you can use an openai api key with cursor. Astra is just not included in the cursor plan.

What does this question have to do with the linked article?

I keep seeing mention of the cache, what's special about it? All frontier llms have prefix caching, what is special about deepseek's approach?

Openrouter's pricing via the deepseek provider is the same as the official deepseek api for both flash and pro and for cached and uncached tokens. It's literally the same api.

And no, cache rates are not different if you're going through the official deepseek provider. The only way caching rates can drop is if you let openrouter fully control routing by preferring uptime or something, and then it might bounce you between providers. But you can control which providers for a given model are in its routing pool and stop that.

Sure, but they at least quantified it with data. It's not like they just dropped a sentence saying the above, they showed numbers.

Did you read the blog post? They compare to deepswe and call it out as the worst one for false positives (failed, but the benchmark assessed it as correct). It also has less language variance.

This isn't how LLMs work. They aren't self aware like this, they're trained on the general internet. They might have some pointers to documentation for certain cases, but they generally aren't going to have specialized knowledge of themselves embedded within. Claude code has no need to know about its own internal programming, the core loop is just javascript code.

This take sucks. The anticheat software in this context is for competitive games. No one cares about people cheating in isolation in single player games. The anticheat is to stop 1 guy from ruining it for the 9 others he's playing with online.

You can argue about the methods used for anticheat, but your comment here is trying to defend the right to cheat in online games with other people. Just no.

Honeycomb is inspired by Facebook's Scuba (https://research.facebook.com/publications/scuba-diving-into...). The paper is from 2013, predating honeycomb. Charity worked there as well, but presumably was not part of the initial implementation given the timing.

MCP is just function calls with parameters. Whether or not it's push or pull can be decided by the author. A push model takes the scan as an input to the mcp call. A pull model does the pulling within the mcp call. Neither is right or wrong, it's situational.

To me, dabbled means learned the fundamentals of the language but hasn't written real software in it. I would expect a dabbler to have been exposed to class inheritance and primitives such ss symbols.

That makes sense as a general pattern, but probably less so in the context of rails which is typically a self contained monolith. It would be adding another hop, more indirection, and more complexity. It would introduce new problems like the need to segment your "real" api from your worker api for the purpose of load isolation.

How does this reduce operational overhead? You still need a queue and a worker to dispatch the tasks to your api.

Your gemini pricing is for flash, not pro. Also, claude uses prompt caching and gemini currently does not. The pricing isn't super straightforward because of that.

They emailed us in advance of the update being pushed out saying it would lower battery life, and offered compensation options ahead of time. It wasn't reactionary. If it were a buggy update the email wouldn't come before the update.

Like I said, they emailed us ahead of time and THEN pushed the update. Before the update was pushed, they said the battery life would deplete and offered compensation options.

They haven't said why they did this, like if it was a battery defect issue or what.

This wasn't a buggy update. It was announced ahead of time via email, including compensation options.

The economy has been "about to take a nosedive" for quite some time now.

Good news, red reader was granted an accessibility exemption so it's sticking around.

This is such weird commentary that pops up on every algotrading post. Why does it have to be socially productive? No one makes the claim it is.

This seems like a decision passed down from the card networks through stripe. Multiple gateways wouldn't necessarily help here.

It's in stripe's best interest to fight the fine. Why would you think they would prefer to pay the networks and collect the fine from a merchant?

This also means an attacker can be running around with a compromised token for up to a half hour before they're stopped.

I'd imagine they do this to not tip off spammers when a message goes through. It's the same idea behind returning a 404 when trying to access a resource you don't have permission to, or not telling users if an account actually exists under an email when doing a password reset.

This is a technical post, stay on topic and stop posting flamewar bait.

This comparison misses the point. Go requires you to bubble up errors manually, that's the difference.

Native static binaries are such a small part of what defines Go. Go is absolutely a competitor in the jvm/.net space. C# and Java are its closest competitors since they're all performant, statically typed, and garbage collected.

Compiling to binaries is an advantage but I can guarantee it's not why most people use Go.

The context here is whether or not they're still happy with it.

Could you use EFS to host a horizontally scalable relational database by using finely sharded SQLite dbs? Like if you had 1 db per user, for example.

Because the ruby community generally revolves around rails. Stack overflow often has generic ruby questions where answers assume you're using rails (by referencing activesupport, for example). Just to give you an idea, on stackoverflow, there are 222k ruby questions and 330k rails questions tagged.