robeym

So I'm guessing as a result of this, FAANG and similar companies are going to have free access while the little guy will be left out.

Is there any concrete information on whether this would be an application process or perhaps for companies that meet certain criteria?

I'm on 0.139.0 on Linux and my visible ~/.codex/logs are only about 129 MB.

This makes me even more conservative on upgrading these tools every time they prompt me to. Better to let them get a few miles on them and see how the community responds.

In this case it sounds like 0.142.0 reduced the issue but didn't fully settle it. I'll wait for 0.143.0+ and see if that version is more acceptable.

My opinion is that the publicity can only help Cursor. I don't necessarily think SpaceX would make Cursor better. Copilot (which I view as a direct competitor to Cursor) has a huge structural advantage. I have several friends in various American companies where Microsoft products are all they are allowed to use. They get "free" Copilot access as a part of their Microsoft plans. Developers aren't having Cursor placed right in front of them in the same way Copilot is, and from my experience, when developers have the choice to pick one, they pick Cursor. So, I just feel the SpaceX/xAI publicity could help Cursor get more visibility in these general American software companies more so than they could on their own.

Cursor was my first hands-on experience with AI. I didn't know much about getting set up with specific providers via API, and Cursor made it easy to pick any model, ask a question about some code, and get a clear suggested answer easily viewable in the IDE with an 'accept' or 'reject' button. I think they answered this question well: "How do normal developers want to interact with AI?"

I moved away from Cursor when I noticed the responses from specific models were not as clean or accurate as when I'd prompt the models directly, which was something I didn't know how to do early on. I hypothesized that they had some boilerplate prompt sitting atop of my own, causing less precise or desirable results.

I would assume Cursor is still one of the best options for normal developers to get started with AI, but with Copilot forcing their foot in the door at many companies, I wasn't sure how well it would fare on its own. Being acquired by SpaceX should help, and I'll be interested to follow along and see how things develop.

Great post. It was an amusing read, and quite the discovery! Good work, and great job documenting it.

Unfortunate to see AI police as top comment on a good amusing post!

Given how easy it was to get our own version of an AI service tool set up, I'm shocked to see a huge, capable company buying it's way into that functionality as opposed to using their large existing datasets to fine-tune an internal tool. Is 3.6B really the better option?

Ah yes the bait and switch. Didn't take long. Win some market share from unhappy AWS customers, hike the prices later

I think it felt more polished, otherwise they wouldn't have released it. Fable took more shortcuts, lied more, had stabilizing safeguards removed, and just did what it wanted more than other models.

PAX ERP. An AI-assisted ERP for small regulated/job shop manufacturers that have outgrown QuickBooks and can't afford (or take the time) to go through a traditional ERP implementation.

PAX is the easiest ERP to pick up. Our core idea is that ERP should not take weeks of formal training and implementation. We have no formal training, no implementation fees, and are a complete ERP+CRM GAAP accounting system. It's intuitive, well-documented, AI-assisted (never steered), and people can pick it up and start using it well with no questions to our team.

Having used and watched others use Fable, it's clear to me that the standard of care for these models done by Anthropic pre-release was nothing like their previous releases. It seemed like shortcuts were taken to get it released. Maybe the models are changing faster than their team can adapt, and maybe the government responded as quickly as they did because of their relationship with Anthropic. Regardless, I think regulation was appropriate here. Anthropic can take the time to refine and improve their internal models to have a more successful release next time. It is normal for the government to restrict potentially dangerous companies if the level of care seems below what is acceptable in the US. We've seen it a lot recently with FAA grounding rocket companies. I don't feel it is cruel treatment just cause it's Anthropic

It's good to see companies thoroughly testing these models in ways that could be used against them, but this should've been something Anthropic did more of before releasing it in the first place. The standard they used for releasing Fable seems much lower than previous releases.

The skill floor for attackers has collapsed, and I think regulation against Anthropic is appropriate here - as much as I am generally against regulation!

Fable has been Anthropic's most ambitious and hopeful release. It makes me think Mythos isn't anything but Opus with certain guardrails removed. Very interesting. Hoping we'll see some quick refinements to it

It's been amusing to watch the AI trend of increasing unusual tool uses. Fable easily takes the cake. I learn a lot more terminal commands thanks to it!

For PAX ERP. It'll be read by both humans and llms. "Paxy AI" will be the primary reader, which is the AI support tool in the system. When users ask it a question about how to do something, it'll answer based on the technical docs.

AI impacts so much. So many white collar chores just get obliterated with AI. Legal & regulatory docs, production planning, sales materials, etc. Just yesterday I used AI to generate 235 new system docs based on our codebase, and added automated .md -> html publishing so final drafts go straight to the website. This work that would've taken a contractor 1-2 weeks got done in a day, by me, someone who definitely isn't a technical writer.

When the person who knows what's needed can handle the technical execution themselves, you no longer need that second person.

I certainly wouldn't say you need "more humans"

That woodworking idea is great. It's just a hobby for me as well but after long days of desk work where moving fast is rewarded, it's great to work slowly on a wood project with my hands. I tried listing a couple things on Facebook marketplace but I had never used it before so they just removed my posts.

Fun read! This is a good story to share. I spent over a year building a complex SaaS product for a narrow market, but it was something I personally needed and was interested in. It was a second full-time job as a solo developer, but the chances are if you are interested in something and find it worthwhile, others will too. It's worth it! Great work. Fun story. Thanks for sharing

Pull Requests, Issues, Git Operations and API Requests

I think quality experience and personal values are more important than engineer cost. I've seen so many shortcuts taken on outsourced work these past few years. AI also loves shortcuts. The combination of the two is not worth the cost savings.

If you value high quality work and pride in what you do, outsourced workers (who most often don't pay careful attention to their work, hence the cost) are not the solution. However, if you're just trying to get something done and don't care about it getting done right, what better way to do it than spending the least amount of money possible

I don’t think we should be discouraging people from buying homes. In a healthy market, the goal should be more households owning the homes they live in, not more homes being accumulated by investors and non-owner-occupant buyers.

The mistake is not buying a home. The mistake is buying too much home, stretching the term too far, and ignoring maintenance.

A cheaper house, shorter loan term, and realistic repair/addition budget change the math a lot.

Own what you can actually afford to maintain.

I was in the same boat, moving my max subscription to Codex instead of Claude, with a Claude Design project going. I was under the impression that Design was pro plans only, so I downloaded basically everything before cancelling.

It could be worth a quick $20 subscription just to grab your stuff, then cancelling. Trying to get support from either Claude or OpenAI seems pretty hopeless. Hopefully this post will get them to see you

I think progressing as humans is something to be proud of. I care less about who gets credit and more about what we can now do.

I also do not think this makes people less capable of solving hard problems. The bar just moves up. More people can now work on harder problems with better tools.

If the goal is credit or proving real skill, then focus on harder problems, like ones AI can't reach

I believe this is in response to PocketOS. When I read the original post, I was trying to figure out how they even built a workflow that had AI so close to the self-destruct button. This post's explanation about it probably being fully vibe-coded makes sense. How else would the system be so fragile and for the agent to have such far reach? They built a house of cards.

A far better approach is being precise with your prompts and if you find a new model has any bad habits, address it specifically in your AGENTS.md and go on your way. If you want to throw in slop promts, go ahead and add a massive AGENTS.md your employer gave you.

People waste too much time on this stuff. The next version could totally change how the model processes your agents.md.

Get good at promting, use agents.md as a minimal model annoyance fixer, and reset it often (every major release)

Anytime Microsoft buys something it's time to start looking for alternatives

Most people run to GitLab or Codeberg. The real loss is network effect of everyone being on GitHub.

GitLab: most similar to GitHub. Repos, CI/CD, issues. Hosted or self-host

Codeberg: free, non-profit, FOSS-friendly. Runs Forgejo. Where Zig went

Forgejo / Gitea: lightweight, self-host on a small server

Bitbucket: fine if you're already in Atlassian/Jira land

Sourcehut: minimalist, email-patch workflow. Fast but a culture shift

Radicle: peer-to-peer, no central server. Niche, principled

Tangled: newer, built on ATProto

Solid and growing impressions but low CTR. We're not quite on page 1 for most queries, and not even close for our primary target queries.

I've used Claude Code from the beginning and the first waves of changes were genuine improvements, with a very steady 4 months in late 2025. But recently, these past 2-3 months, the changes have shifted towards significant and frequently degradations in model performance. I had a very consistent workflow for 4+ months in Claude Code, but this year I've had way more surprises. I'm not sure which tools you use, but Claude Code had a great period of consistency, at least compared to other AI tools.

As a long-term 20x user, Claude has recently felt a lot like using AI for coding a year or so ago. It can't reliably handle basic tasks. I ask for something straightforward and get something subtly wrong, incomplete, or just not workable. I always use the best model available and effort levels maxed, but with all their changes I have to relearn how to make the model perform at best every day, and it seems I can't keep up. It’s not that Claude can’t do impressive things, it clearly can, but the inconsistency on simple, expected behavior makes it hard to use. The downtime is annoying but hasn't been the deciding factor. I’m not waiting it out this time. I’m switching over to Codex, and based on my usage today it looks like I’ll be fine on the 5x plan, so I can drop down and save about $100 a month which is nice. I didn't quite have a grasp on how quickly companies can change for better or worse until Anthropic showed me. I'm surprised at how quickly they brought me from a happily paying max user to not even wanting the lowest paid tiers.