Hacker News Reader

spaceywillyon Claude Sonnet will ship in Xcode

I use VS Code with Claude Code, then I just use Xcode to build and launch

spaceywillyon Waymo granted permit to begin testing in New York City

Honestly the train system in NYC is so good, I have only taken a cab a few times since I moved here. I’ll probably take a waymo once if they roll it out here for the novelty of it, but I’d rather see people getting exciting about public transit. Life is so much better when you don’t have to depend on cars to get you places.

spaceywillyon AI tooling must be disclosed for contributions

As long as they make it easy to add a “made with AI” tag to the PR, it seems like there’s really no downside. I personally can’t imagine why someone would want to hide the fact they used AI. A contractor would not try to hide that they used an excavator to dig a hole instead of a shovel.

spaceywillyon Claude Sonnet 4 now supports 1M tokens of context

I’ve had more success this way as well. I will use the model via web ui, paste in the relevant code, and ask it to implement something. It spits out the code, I copy it back into the ide, and build. I tried Claude Code but I find it goes off the rails too easily. I like the chat through the UI because it explains what it’s doing like a senior engineer would

spaceywillyon Duolingo CEO tries to walk back AI-first comments, fails

Not to mention, at that point they are essentially just serving as the middle man between you and a LLM. Instead of paying for Duolingo to ask an LLM to generate a bunch of Spanish phrases for you, you can pay for the LLM and do that yourself plus a lot more. I’m not sure Duolingo has a solid understanding of why they exist as a business.

spaceywillyon Claude's system prompt is over 24k tokens with tools

Yeah it’s kind of like we have invented a car that drives around wildly in any direction, and we are trying to steer it by putting up guard rails to get it to go where we want. What we need is to invent the steering wheel and brake pedals, which I’m sure smart people are working on. We’re just at a very early point with this technology, which I think people tend to forget.

spaceywillyon TScale – Distributed training on consumer GPUs

> and I struggle to think of what would lead one to the urge to implement a key=value config file parser in 2025

That could be a symptom of LLM coding. I have found at times they will go down a rabbit hole of coding up a complicated solution to something when I know that a library already exists it could’ve used. I’m sure part of the problem is that it isn’t able to search for libraries to solve problems, so if its training data didn’t use a particular library, it will not be able to use it.

spaceywillyon Vibe Coding is not an excuse for low-quality work

From that tweet:

> I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it

That sums up vibe coding, imo.

The article talks about code quality with vibe coding, but I think that misses the point. The real problem is code knowledge. When a vibe coder inevitably needs to debug something, if they have no idea what any of the code does, or why it is the way it is, they are not going to have a good time.

Sure they can copy paste the error into the LLM and hope for the best, but what happens when that doesn’t fix it? I’ve already had to spend hours at work tracking down bugs that ending up being in there because someone just blindly accepted the code an LLM wrote, I fear it’s only going to get worse.

spaceywillyon OpenAI o3 and o4-mini

They have 500M weekly users now. I would say that counts as doing something.

spaceywillyon OpenAI in Talks to Buy Windsurf for About $3B

To play devils advocate, software engineering is currently proving to be a hugely valuable use case for AI, and will almost certainly grow over time. I’m not sure what the market cap for software engineering tooling is, but I assume $5B or even $10B would only represent a tiny fraction of the TAM.

Here is one report that projects a $25B market by 2030. It would not take long to recoup their investment if that prediction comes true.

https://www.globenewswire.com/news-release/2025/03/26/304970...

spaceywillyon A Reddit bot drove me insane

https://en.wikipedia.org/wiki/Dead_Internet_theory

- sent from a human

spaceywillyon Meta got caught gaming AI benchmarks

I agree. I think of it like a car engine. You can push it up to a certain RPM and it will keep making more and more power. Above that RPM, the engine starts to produce less power and eventually blows a gasket.

I think the performance-based management worked for a while because there were some gains to be had by pushing people harder. However, they’ve gone past that and are now pushing people too hard and getting worse results. Every machine has its operating limits and an area where it operates most efficiently. A company is no different.

spaceywillyon What If We Made Advertising Illegal?

“What if we made advertising illegal?”

Many small companies would go out of business, that’s what. Yes we definitely need advertising reform, but advertising is a very important part of any business if they want to be successful. Making it illegal would cut a lot of businesses off from their potential customers. The author doesn’t seem to propose any alternative solution for this.

spaceywillyon Why I'm No Longer Talking to Architects About Microservices

“In the end most systems will work somehow”

The older I get the more I feel like this is a feature, not a bug. I have spent too much time over engineering systems that end up being canned because leadership decides to go in a different direction. These days I am much more concerned with striking a balance between perfect and good enough.

Get something out into production, start generating some revenue and finding the stuff people like about it, then see what is giving you the most problems and fix that. That’s going to be a much more successful path than never shipping anything because of endless polishing and optimizing.

If I ever find myself spending a long time debating some technical decision (aka bike shedding), I have learned to hit pause and say let’s just pick a path and go with it. As an aside this also works for deciding what restaurant to go to or what kind of pizza to order.

spaceywillyon Whose code am I running in GitHub Actions?

This does not bode well for genetic AI

spaceywillyon Microsoft is plotting a future without OpenAI

I actually think the AI is going to end up creating its own sort of machine code. Programming will be done entirely in natural language, the AI will translate to machine code and we tiny brained humans won’t even know or care what it’s doing under the hood. The idea of programming using a specific programming language is going to seem archaic and foolish.

spaceywillyon IPv6 Is Hard

Yes. In my day job I work mainly on containers running entirely in data centers we own, they are fully ipv6, it’s very nice to work with.

Then there is this one customer, they are colocated in other ISP owned datacenters, and some of their server maybe have fully IPv6, some may have only IPv4, some may be a mixture of both. Supporting them is a never ending nightmare.

spaceywillyon Our channel on YouTube has been deleted due to “spam and deceptive policies”

We need decentralized services that are not subject to the whims of whichever power hungry billionaire is in charge this week

spaceywillyon Introducing deep research

I think a lot about how differentiating facts and quality content is like differentiating signal from noise in electronics. The signal to noise ratio on many online platforms was already quite low. Tools like this will absolutely add more noise, and arguably the nature of the tools themselves make it harder to separate the noise.

I think this is a real problem for these AI tools. If you can’t separate the signal from the noise, it doesn’t provide any real value, like an out of range FM radio station.

spaceywillyon The 'Return to Office' Lies

I resisted RTO as much as anyone but eventually I took a job that required me to be in office 3 days a week because, well, didn’t have much of a choice.

I have to say being in the office has been better than I expected. I recently got together with a group of former coworkers, who I worked with for many years at a startup, and have kept in touch all these years later. We noted that this kind of friendship definitely would not have happened if we were working remote.

My only thing is, I would rather have the flexibility to choose where I work, even if only at the team level. It gets cold where I live in winter, I would love to be able to go stay someplace warm during those months, and come to the office when I want to be there.

I feel like for me, this debate isn’t really about working in the office vs working from home, it’s about control. Companies have realized that they gave over too much soft power to employees during the pandemic, so they are now working together to claw it back. They could care less where we work, as long as they are the ones in control.

spaceywillyon Phased Array Microphone (2023)

This is known as the Cocktail Party Problem. It turns out or brains do an incredible amount of processing to allow us to understand a person talking to us in a noisy room.

https://en.wikipedia.org/wiki/Cocktail_party_effect?wprov=sf...

spaceywillyon Netflix buffering issues: Boxing fans complain about Jake Paul vs. Mike Tyson

To me the difference is that in 2012, you had companies focusing on delivering a quality product, whether it made money or not. Today, the economic environment has shifted a lot and companies are trying to increase profits while cutting costs. The result is inevitably a decline in quality. I'm sure that Netflix could deliver a flawless live stream to millions of viewers, but the question is can they do it while making a profit that Wall Street is happy with. Apparently not.

spaceywillyon No GPS required: our app can now locate underground trains

I take the train every day in NYC. Around half the time the train is brand new and has nice displays that even show you where the stairs are located in each station as you pull up. The other half the train is from the 1960s and has no information whatsoever. So yeah this app will definitely be an upgrade over peering out the window to read the station names. Even the trains with displays are occasionally wrong/not working.

spaceywillyon DeskPad – A virtual monitor for screen sharing

I use the “maximize video” chrome extension which may work for you. You can click on any video player and it will make that take up the whole browser window size. So then the video size == the browser window size. I use it to panel multiple videos around my screen (mostly for watching multiple NFL games at the same time).

I also use Better Touch Tool which supports keyboard shortcuts for arranging windows, I believe there’s a similar tool for windows. So for example if I want 4 equal sized windows (in each quadrant of the monitor) I can do it easily with keyboard shortcuts.

spaceywillyon Our container platform is in production. It has GPUs. Here's an early look

Amazon has had serverless functions for a long time now. I built an iOS app with a backend in AWS and it was as you say, “here’s my code, you make sure it’s running somewhere.” I uploaded my Typescript code, set up an API gateway to call the lambda function and… that’s it. No load balancer, no ECS management. It’s been running for years and I haven’t to do anything other than pay the bill every month.

spaceywillyon So you want to build an embedded Linux system?

Can you check if there is another USB port exposed via a pin header? That is probably the easiest way if it is available. Otherwise, I think your best best is to solder wires onto the same pads used by the existing USB port. I wouldn’t try to desolder the port that is there, that could cause more problems. But if you can find a location on the board to tap into that USB port’s signals, you could solder wires to run another usb port using the same connections.

spaceywillyon Ditch the sim rig and use your car instead (2019)

BRZ has electric power steering assist. If he puts the wheels on alignment pads he should be able to move the wheel easily. I’d still be worried about wear and tear on the components (you really don’t want your tie rods to give out at speed) but it’s a fun weekend project.

spaceywillyon Launch HN: Airhart Aeronautics (YC S22) – A modern personal airplane

I think battery electric planes will make this less of an issues, once they arrive. The motors are quieter, vibrate less, don’t generate smelly fumes, and there are more mounting options. The eVTOLs from Joby are supposedly very quiet in operation.

spaceywillyon Launch HN: Airhart Aeronautics (YC S22) – A modern personal airplane

I worked for a company doing something similar and they used servos that were strong enough to control the plane, but no so strong that the safety pilot couldn’t overpower them if needed. I think it’s the same for any autopilot really. And there’s always a big “disconnect autopilot” button in case it goes haywire.

I think they will need to consider that situation though, if the autopilot does fail the pilot would still need to have enough skills to land the plane. Even ATP pilots have this issue where they over rely on automation too much and lose their ability to land manually in an emergency situation.

spaceywillyon Launch HN: Airhart Aeronautics (YC S22) – A modern personal airplane

You will also need to account for the time to pre-flight the plane, and fuel it up, possibly wait on the taxiway, and possibly divert around weather (including to an entirely different airport if necessary).