mbreese

Capture log messages and write them directly to stderr? That’s easy.

This is where the LLM coded library replacements make sense. When you have non-critical code that can be easily swapped, I’m happy to let an LLM generate a logging package. If I have a need to send log messages out to a separate server for archival? I’d be a little more critical.

Plus, hams have a lot of geek history, so anything that’s about ham radio is probably going to get votes. I’m sure the time of day doesn’t hurt either. This is hitting at a good time of day for a US based post like this.

I’m not convinced that the economics are favorable to space based data centers, but ironically, it still might be a lower hurdle than terrestrial data center approval.

It’s all about segmentation of the market.

Apple has customers that want a small, well designed all in one computer that fits on a desk. If it has a pop of color, all the better. It is a customer segment that Apple knows well and has been serving since the first Macs.

There are many companies that want to address the market for a budget friendly small PC with an external monitor. There aren’t many that try to address this segment of the market.

I can picture that they’d have the software side sorted out with RedHat. But I don’t know what hardware they’d push to support local on-premises models. It looks like when you buy Fusion HCI, you can choose from Dell or Lenovo as the underlying hardware supplier.

This is more of a statement for their consulting customers, right?

I think that's the difference here -- OpenAI and Anthropic are competing more with each other (and Google Gemini) than they are with open weight models. While people on HN are familiar with Kimi and GLM models, I'm not sure how many casual users of LLMs know about them. I think the competitive risk open weight models pose to OpenAI and Anthropic is much further in the future.

There is a lot of usage that's only known by Anthropic or OpenAI that's not visible to the rest of the world.

I know which I’d expect to perform better.

Or did I miss something?

Non-technical people believing that these models are actually thinking can lead to all sorts of inappropriate uses.

For most of us though, you're right -- the OpenRouter option is far more cost effective.

Of course, this would also make it easier for someone to disable a speaker and track a person without their knowledge, making it more difficult for that person to find a stealthy AirTag.

I'm not surprised that someone didn't think this through. I am surprised at how many people this must have gotten by to make it into production. You'd think that with any AirTag related product, privacy would be more central to the planning.

Siri, text my dad … send it. Thank you.

And then it could stop trying to listen. It sounds like you’re being polite, but it would have a secondary meaning too. Kind of like using the stop word to end an LLM’s output token list.

This perspective creates so many unanswerable questions.

Actually, I’d really like to know the answer to this. I’m not a fan of clickable TUIs. Mouse events in my terminal are one of the banes of my existence.

But, if they are more popular with others, that would be helpful to know.

We’ve had low quality code for as long as we’ve had programmable machines. I’ve seen a lot of manually written bad code well before LLMs took over. Nowadays, I primarily look at how long a repo has been around and how active it is. A repo that has consistent activity is probably going to be a higher quality one. Even better if I can follow PRs to see how the code base is evolving.

I will admit it’s become more difficult to tell low quality git repos from higher quality ones. Missing files or incomplete README files are a thing of the past. Claude is very good at filling in the gaps to make all repos at least appear respectable.

Why pay to reach many thousands of people with a widely read paper when you can pay pennies to reach dozens of people that you don't want to apply to the job anyway?

This is an entire business model.

It failed.

But I think that they learned something really valuable. As a company, they don’t have a good understanding the server market or how to sell to it. Their premium product style just didn’t have a place in the race to the bottom server market.

We had two in my grad school lab that we used as web servers. But that was the problem - it was a premium designed product that wasn’t any better of a web server than a plain (cheaper) Dell 1U box.

Maybe things would be different now that they have a really unique chip. ARM as a whole ecosystem has more buy-in from other companies than PowerPC did. Their performance per watt and GPU design might make more of a difference this time around.

But I can’t see Apple being able to release and sell a data center centric server. I think others might have a better chance.

I do find it interesting though, that in the interest of transparency, every bugfix gets a CVE. Which ends up being a huge number… which will ultimately yield a more insecure environment as we’re getting conditioned to ignore/discount CVEs by the volume.

Over-reporting in this case seems to risk being counterproductive.

If NASA was having visible issues and difficulty in trying to get something like Hubble working in space, then perhaps the US’s spy agencies (known or not) also had the same difficulties. NASA employees wouldn’t need to have known about the NRO to help provide cover for their capabilities. Just publicly showing their struggle - for transparency - would have been enough.

(If NASA had a clue about the NRO capabilities or not, I doubt they would have cared. They still got their space telescope!)

This is more like alcohol apathy - not drinking, but also not explicitly anti-alcohol. That the more generational shift in Gen Z that I've seen. But, those sentiments can certainly change over time.

My personal theory is that they are scared of ending up on the Internet when they don't have control.

Even in the case where it’s an internal only Lu array, it can be complicated to refactor a library name.

I was going to make a comment earlier about how my kids travelled in car seats years ago before I realized you were talking about lap kids without their own seat.

The phased array antenna is also a function of this. Each satellite is moving so much that you need to have an antenna following the satellite across the sky.

So, once Starlink decided to have a low earth orbiting constellation (for better latency and bandwidth), the other features (quantity and a steerable antenna) became requirements to have a functional system.

1) you send a signal to the satellite

2) the satellite relays the signal down to a ground station

3) from here your packets go out on the Internet like any other ISP

The answer then goes in reverse: back to the ground station, up to the satellite and down to your dish.

Starlink’s special sauce is the low earth orbit. Older satellite internet companies operated in geosynchronous orbit. This meant they only needed one or two satellites to cover a large area. You also only needed a small number of ground stations. The satellite needs to be able to see your dish and the ground station at the same time. But this is very high up, so latency was horrible (the distance up and down before the packet even hits the internet). Bandwidth was also limited, but much less of an issue.

With low earth orbits, you have great latency. But, because you still need to be able to see the ground station and the customer’s dish, this means that you need more ground stations. This is just geometry.

Starlink adds one more feature that is relevant here - direct satellite to satellite communication. Instead of the signal going from:

Customer - satellite - ground station - internet

The signal can go from:

customer - satellite A - satellite B - ground station - internet.

This later setup might be what you were thinking about. You still want to limit the number of hops from satellite to satellite you have (latency again), but this should make the service more resilient to the loss of a single ground station.

For a marketing page, it’s not the worst UX I’ve seen, but still slightly annoying.