Lot of things going on this summer, but yesterday’s big news was the release of ChatGPT 5. I don’t have an opinion on it yet, but I guess it’s nice to have a single model (at least exposed, even if this is routed under the hood) instead of switching between 4o and o4-mini-high, etc.
Recently I’ve been on the market for a “CLI Agent” (synchronous, like Claude Code, not async like Devin and co.). I’ve been using Gemini CLI with good results — as long as I’m within the free PRO requests. As soon as it switches to the “flash” model it becomes useless, can’t edit a file properly, and just loops on itself. I wouldn’t pay for it. Claude Code seems too expensive; I might give OpenAI Codex a try with GPT-5, but I haven’t heard much about it.
I’ve been vibe coding a “sun earth moon” system with day and night cycles. Nothing fancy, it’s to teach my 3-year-old about space, time zones, and moon cycles, and it’s fun to make something graphical. Eventually I want to get the moon cycles right (ideally in sync with the real time on Earth).
Let’s see, what else? Also on Google’s side, I noticed I had access to the video diffusion model Veo 3 in the Gemini app! It’s pretty good — you can prompt from scratch and also provide a starting image for the video. I got this video of Jensen plugging a GPU in a rack. The stop-motion is a nice touch for a Lego animation.
Lastly, I wanted to speak about Open Source. This week there was a popular post on Hacker News that I found pretty tasteless, I gave the AI arms and legs then it rejected me. The short story is that the author found out that Anthropic uses his open source library as part of Claude Code and he somehow thought that ought to give him a chance for a job interview, and when that didn’t happen (likely an automated rejection) he went to blog and post about it (a second time in Ask HN: Have you ever regretted open-sourcing something?).
My take is that if the library was that valuable and complex, maybe don’t give it away for free in the first place? I’ve considered for a long time now that Open Source is devaluing software engineers by taking away the “fun parts” (technically interesting) and just leaving the dregs of plugging stuff together. As a software engineer working on a new feature, you’re usually expected to find a suitable existing library rather than code your own, so you end up with a very shallow knowledge of the subject. Don’t get me wrong, I’m glad for Redis, Nginx, Linux, and countless other software etc. being open source! But notice that they are backed by companies or sponsors; it’s not about an individual or two toiling away and complaining about not being funded. The reason for using an open source library is that it’s free — you can’t expect to profit from it. See what happens when popular projects like Redis actually tried to change their license — they got forked immediately.
Anyway, I tend to favor languages that come with a comprehensive standard library — see how in Go you usually make web apps without needing a framework — and reduce the number of vendored dependencies along with the security risk. And I don’t see any reason to make my own projects open source (save for one-off scripts etc. that can be useful to some and wouldn’t make money anyway).