mmoustafa

I like how they called it “ML Infrastructure” instead of “AI Datacenters”

Nothing is new, quick tunnels were launched in 2021

https://blog.cloudflare.com/quick-tunnels-anytime-anywhere/

Quick tunnels have existed since at least 2022 (when I was using them)

honestly just Bend is a great HN title, you can describe it more concretely on the homepage

Thanks for the response Chris, I couldn't make my business work without OpenRouter in the first place, so kudos

This was meant as more of a technical reference, sorry you had to wake up to a PR drill lol

Yes, you hit it on the head. It's worth it, but you have to do your own evaluation and promotion.

Example I forgot to mention: `:nitro` ranking is not the fastest, I do a round robin sampling with representative payloads to find out the fastest providers and reorder my list on the fly.

I'm confused, what do they mean when they say they reduced prices?

DeepSeek v4 flash is $0.10 / $0.25 as opposed to this v4.1 bump which is $0.30 / $1.20

I haven’t reverse-engineered the GrokBot comms layer but it’s pretty similar in experience

not impossible, pypush did it 3 years ago

Yes, it is awfully inconsistent today. They run some tests (accuracy table buried on model page) but they are sparse and only capture a single moment in time. I would love to see OpenRouter take this more seriously.

It is just the most visible “close call”. Most of the time we don’t hear about these incidents until they blow up, that is if we hear at all.

How do you define vibecoding? Is using LLM-assisted autocomplete vibecoding? Is editing single files via an LLM vibecoding? When do we cross the threshold?

Just saying mostly written by generative AI is extremely broad as that can happen even with “human written” codebases via AI-assisted tools.

The out-of-stock $6000 M3 Ultra Mac Studios with 800GB+ memory bandwidth are going for $24,000 on eBay, so yes definitely

feel free to (and you should) audit! https://www.npmjs.com/package/@0xmmo/crew

I think you're looking at $60 an hour which is not horrible (estimating Waymo avg is 3 rides / hour x $20 per ride)

how do you select text?

Aaand that would get challenged in court, remember they had to get Congress to create this designation in the first place because it is not de-facto legal for the USG to discriminate between individuals or corporations.

The Supply Chain Risk label requires every single company in the supply chain of a product or service provided to the US Government to either drop Anthropic or get dropped themselves. This is not just suppliers, but also includes suppliers of suppliers all the way down. This is a much larger chunk of the economy (approaching 100%) than the Pentagon/DOW.

I would love to see real-life tokens/sec values advertised for one or various specific open source models.

I'm currently shopping for offline hardware and it is very hard to estimate the performance I will get before dropping $12K, and would love to have a baseline that I can at least always get e.g. 40 tok/s running GPT-OSS-120B using Ollama on Ubuntu out of the box.

Why do all the discussion posts about ICE’s biometric app get taken down? Although they may invite politicing, they are very relevant to HN.

e.g [flagged] Target director's Global Entry was revoked after ICE used app to scan her face [https://news.ycombinator.com/item?id=46833871]

They do not at any point outline how cooling will be done, they simply say "it will be more efficient than chillers due to the larger delta T" which is incorrect because it's about dT not delta T

> "The only logical solution therefore is to transport these resource-intensive efforts to a location with vast power and space. I mean, space is called 'space' for a reason. [crying laughing emoji]"

This is all the reasoning provided. It is quite sad how a company I admired so much has become embroiled in financial doohickery.

It is not about tracking changes, git does that well enough.

It is about tracking the reason for the changes, i.e. git messages but on steroids.

new trolley problem just dropped: save 1 billion people or ...

I wrote an article on this exact issue (albeit more simpleminded) and I suggested a rudimentary way of tracking provenance in today's agents with "reasoning traces" on the objects they modify.

Would love people's thoughts on this: https://0xmmo.notion.site/Preventing-agent-doom-loops-with-p...

Although this is what Opus recommends, it will give you many issues as you don’t really have any node runtimes in the path (or worse if you do).

What I recommend is replacing it with $PATH=(a command to find the nvm default alias directory, detect the verion and load it from that specific version directory directly) so you always have default node in path and then lazy loading only nvm itself, so you can switch when you need to.

Sorry I don’t have the command handy as I’m on mobile but if you paste the above into Opus you’ll get it.

I don't think 0.38s is a bad trade-off for convenience when the rest of the tools I need to do my job collectively are another 2s at shell startup. NVM alone adds 0.5-0.6s on my M4 Macbook Air.

Bigger fish to fry if we're being practical.

I did, it's available at olly.bot :)

> All encryption is end-to-end, if you’re not picky about the ends.

This reminds of how Apple iMessage is E2E encrypted, but Apple runs on-device content detection that pings their servers, which you can't possibly even think of disabling. [1][2]

[1] https://sneak.berlin/20230115/macos-scans-your-local-files-n... [2] Investigation in Beeper/PyPush discord for iMessage spam blocking

I built a connection to a web-powered LLM over SMS/iMessage for literally this purpose. While traveling I’d have really bad or sparse service but still needed to find my way around.