Hacker News Reader

pseudosavanton OpenAI begins rolling out GPT-6 Astra

If the numerous benchmarks in that post are to be believed, Astra is really next level. I can't wait to try it out!

pseudosavanton Paint.net 5.2 alpha now runs on Linux

English only has "free". PDN is gratis, free as in beer. While the free software platforms (WINE, Linux) are libre, free as in speech.

I still think the argument holds though that people who have maybe used PDN but haven't contributed time, code, money, or even bug reports have no business raging at its author because he used AI to make it run on WINE/Linux. Literal ingrates.

pseudosavanton The ChatGPT/Codex app bundles a full copy of LibreOffice

For those that are interested, this inspired me to create a tool just for exporting images and PDFs of Word, Excel, and PowerPoint files. Check out the repo below, or if you have uv installed just run `uvx office-export skill install` or `uvx office-export --help` to check it out.

It uses Office for Windows, so it is Windows only. I don't currently have access to a Mac machine, so I can't build in Mac support yet.

https://github.com/pseudosavant/office-export

pseudosavanton The ChatGPT/Codex app bundles a full copy of LibreOffice

Cool project! I'm playing around with some similar stuff, but via uv/Python. Impressive you aren't leveraging any other existing libraries for OpenXML parsing, editing, or rendering. Any especially tricky gotchas you've run into for this tool?

pseudosavanton Paint.net 5.2 alpha now runs on Linux

The backlash is so completely unwarranted. This is a piece of free software where the creator is making it easier to use on free software platforms. If you don't want PDN on Linux because you hate AI no matter what, then don't use it. Don't rage attack the creator so that others can't use PDN.

Thanks Rick Brewster! I love PDN. It isn't as capable as Photoshop, but it isn't a rental, and it does 99% of what I need it to do. I've even bought it on the Microsoft Store just to support it.

pseudosavanton The ChatGPT/Codex app bundles a full copy of LibreOffice

This can absolutely be done programmatically. I just added the ability to do that using PowerPoint for Windows yesterday in my uv/python tool `markdown-pptx`. Codex added the functionality in less than 5 minutes.

https://github.com/pseudosavant/markdown-pptx/

pseudosavanton The ChatGPT/Codex app bundles a full copy of LibreOffice

Is that what it is using to render and manipulate MS Office documents? That'd explain the poor rendering of some of my files. Bundling all of LibreOffice seems like a pretty huge dependency.

pseudosavanton A browser-based viewer for Office Open XML documents

This is rendering my files really well!

I'd love to see Codex adopt this for its Office file rendering. I've been working on a markdown-to-pptx tool lately, and certain things like tables or specific master slides don't quite render 100% accurate there. Missing table formatting or completely wrong font sizes. But they look 99% perfect here (I noticed one numbered list that didn't have enough whitespace between the number and the text).

It rendered a really long and complex documentation Word file perfectly though.

pseudosavanton Stopping the smart TV from being used against you

Basically, Windows has been installing monitor drivers for a very long time automatically. Usually a single INF file. Recently, companies like LG have discovered and abused that by instead of just that single INF file, they are installing software that pushes malware like McAffee anti-virus onto your computer.

More of a Windows Update + MS partner issue than EDID.

pseudosavanton The Teaser Period: Why the AI Boom Is Hitting a Reset Wall

Even if it is 5 or 7 years, houses last decades. Even cars last longer than GPUs. GPUs are a highly depreciating asset. They'll last longer than 3 years, but might not be very cost effective (tokens per Wh) compared to the latest AI accelerators that are available.

I wonder if it will be that the U.S. (and allied countries) will be using the latest 1-2 generations of AI accelerators, and if the 3-5+ year old stuff will be sold to Chinese datacenters since that will still be the most powerful tech we'll be allowed to export to them?

pseudosavanton Small Models Have Arrived

It is a good question. Luna is definitely a very capable model. Much more capable than the top SOTA models from 12 months ago. It definitely isn't at the same level as Sol, but you get 20x the tokens for the cost, and it has a much faster tokens/second rate.

If this is a cost conscious company where I'm going to get a fairly limited amount of Sol, or a nearly unlimited amount of Luna, I'm probably choosing Luna.

pseudosavanton The Teaser Period: Why the AI Boom Is Hitting a Reset Wall

1. The massive amounts of GPUs being purchased have far shorter valuable lifespans than a house.

2. A model's value seems to be depreciating at an unbelievable rate. The most expensive top SOTA models (GPT-5, Opus 4.1) a year ago are far less capable than GPT-5.6 Luna. Compared to when those models were new, Luna costs 85% less than GPT-5 and 98% less than Opus 4.1. That's good for us consumers, but if a lab stumbles for 6-12 months, a lot of their value goes away. Especially with open models only months behind the SOTA closed models.

pseudosavanton Meta reaches $17B settlement over social media harms to children

Except that the settlement includes making YouTube and TikTok implement the same stuff. The settlement amount stops by a third if the states don’t make their competition make the same changes.

pseudosavanton Python's pre-declared constants are kinda weird

That is a great xkcd comic! Sums up my feelings well. I honestly can't stand writing Python by hand (whitespace as syntax??) but here I am flying!

pseudosavanton Meta reaches $17B settlement over social media harms to children

Exactly. Meta basically got out of this without any meaningful penalty or deterant for the future. Paid out to the 50 states, it only averages $36M a year per state. I can't believe all of the state AGs signed on for this no-op.

pseudosavanton Python's pre-declared constants are kinda weird

I've never become a fan of the language syntax, but otherwise I've become quite smitten with the total Python ecosystem. The Agents/LLMs + uv combo have made Python so useful and productive for me.

My CLI tools publish from Github to PyPI so that I can run tools with just `uvx sql-agent-cli` or `uvx dlna-here. Nothing for me to handle downloading (directly myself), no environment to manually setup, portable (Linux, Windows, Mac, ARM, x86). Easy for agents to run from a skill.md file without any other prereq than uv.

Really useful library ecosystem to leverage. No more shell scripts, or TS/JS/PHP backend services. I've even used Python on devices I've built around Raspberry Pi Zero 2 boards.

pseudosavanton Dolly Parton has died

Definitely my favorite version of that song!

pseudosavanton The August 17 outage

Not really. In my experience, the average human written commit messages were almost always useless 1-3 word "Should work non" kind of messages. The agents had a low bar to clear.

pseudosavanton The August 17 outage

It’s funny how the quality of the average commit message has gone up, now that we rarely read or write them directly anymore.

pseudosavanton The August 17 outage

For some context, in June they said commits "commits nearly doubled year over year, crossing 1.4 billion per month". Now, it has more than doubled that in just a few months.

https://github.blog/news-insights/product-news/github-copilo...

pseudosavanton HTML Can Do That

There is a common accessibility pattern where the first tab focusable item is a "Skip to content" link in the top left corner. The link is initially hidden until you tab to it. That is the only use case I could think of when checking it out.

pseudosavanton Why the Ocean Cleanup hasn't solved the plastic pollution crisis

Honestly the title of this article is ridiculous. A global problem like plastic pollution, decades in the making, won't be solved by a single organization, especially an NGO, in only 13 years.

I still applaud the word TOC is doing and find it very worthwhile. But let's just be reasonable with our expectations.

pseudosavanton OpenRouter is joining Stripe

Probably a mixed bag at OpenRouter today. Some employees probably just became very wealthy. Others will see what it's like being owned by a different company with different goals. Layoffs tend to accompany most acquisitions within 6-12 months these days. A $7B price tag probably comes with some hefty expectations.

pseudosavanton Rules of good social skills (2025)

I interpret 'breaking rapport' as doing something that makes the other person feel less connected to you. For example, bluntly contradicting them, dismissing something they said, teasing them when they are not comfortable with it, looking away while they are talking, or turning your attention to your phone. Anything that shifts the interaction from "we are connecting" to "there is distance between us."

pseudosavanton Being ambitious and being a dad

I very much agree with this blog post but would reframe it a bit. I think ambition, like most things in life, is zero-sum and requires trade-offs. Before kids, all of your ambition could go into your professional life, but what happens when you become a parent and have parental ambitions?

To have a life that will feel successful to you, you need to have your ambitions aligned with what really matters to you. Whether that is being a great parent, having great community, building a great company, creating amazing works of art, etc.

There isn't necessarily one "right" balance, but there are certainly many examples of people who went all in on one thing (Einstein, Steve Jobs, etc) to the detriment of anything else, even though they brought children into this world. It doesn't necessarily mean they were wrong, but they were probably crap fathers to have.

pseudosavanton Rules of good social skills (2025)

Thanks! The site got hugged to HTTP 500 death.

This is a great list. I didn't disagree with a single point. Some of those lessons took me far too many years to learn myself.

pseudosavanton Remote workers report the highest well-being in study of 7,700 employees

Could agree with this more. I've worked 100% remote for the last 6 years and had quite a few years of hybrid remote work before that.

Not everyone is as well suited to it because you have to implement your own boundaries and discipline. I used to get a majority of my social interaction at work, now I have to make sure I get out with friends, or even just to the store, to make sure I'm not just in my house all day toggling between work and personal life.

But any downside I have to manage like that is more than offset by being at home, with the temperature how I like it (I always froze in offices), with my dogs, able to go pick up my kids from school, and with an ideal desk setup (43" monitor, quiet uninterrupted space, able to use voice mode on Codex without disrupting others) that I couldn't have in an office.

pseudosavanton Cursor launches Origin, GitHub alternative

Not in the same way really. There aren't examples of OpenAI, Anthropic, or GitHub/Microsoft doing the sketchy things that xAI has been involved with. Those companies care a lot about the reputational damage that would come from flouting their customer's data privacy obligations.

Cursor/xAI was making a full copy of every repo it ever touched within the last month, and when caught, just acted like that was an accident. A system to download every repo you touch, and then upload it to persistent cloud storage, doesn't just happen by accident.

pseudosavanton Cursor launches Origin, GitHub alternative

What kind of enforceable guarantees could we actually get that xAI won't use our private code for training? Not empty promises, guarantees that have legal teeth. This applies to anything with xAI or Cursor though, especially after last months' revelation that Cursor was uploading every repo it touched.