Douglas Camata
@douglascamata.bsky.social
I produce code, drink coffee, and run. Working on open source k8s monitoring/Otel. Opinions are my own.
The situation with Github's macOS Intel action runners is awful. It takes 7 minutes from running a `colima start` for the VM to be up and running. :/
November 12, 2025 at 10:49 PM
The situation with Github's macOS Intel action runners is awful. It takes 7 minutes from running a `colima start` for the VM to be up and running. :/
So iOS 26 was buggy. The 26.1 release fixes some things, but is still buggy. When will we be able to upgrade without glaring bugs that are found on day 1? Maybe by iOS 26.3? Apple's quality control used to be better than this.
November 4, 2025 at 6:50 PM
So iOS 26 was buggy. The 26.1 release fixes some things, but is still buggy. When will we be able to upgrade without glaring bugs that are found on day 1? Maybe by iOS 26.3? Apple's quality control used to be better than this.
With the Claude Code Web (and thus mobile) release I see an annoying pattern of people thinking "web" and "mobile" are separate. No, you don't and shouldn't need an "app" to use something on your mobile device. Just open the website and use it. The web is mobile!
October 21, 2025 at 10:06 AM
With the Claude Code Web (and thus mobile) release I see an annoying pattern of people thinking "web" and "mobile" are separate. No, you don't and shouldn't need an "app" to use something on your mobile device. Just open the website and use it. The web is mobile!
I must say that after having to develop some `systemd` services, I started to appreciate how easy it is to configure units and the quality of the docs. The configuration options are good and there are other small tools (Unix philosophy) to help with more complex scenarios. 🤩
October 21, 2025 at 9:52 AM
I must say that after having to develop some `systemd` services, I started to appreciate how easy it is to configure units and the quality of the docs. The configuration options are good and there are other small tools (Unix philosophy) to help with more complex scenarios. 🤩
A tip for working responsibly with AI that I have already seen others talk about but it's worth repeating: you should review any AI generated code just like you would review any other critically important code. AIs often generate overly complex code that can be simplified.
October 17, 2025 at 2:33 PM
A tip for working responsibly with AI that I have already seen others talk about but it's worth repeating: you should review any AI generated code just like you would review any other critically important code. AIs often generate overly complex code that can be simplified.
The year is 2025: can't find a single browser that uses multiple connections to download a file from a direct link. Not even 1 MB/s on the browser, about 30 mins to finish the download (2.5 GB file). Send it to `axel` with 30 connections: 44 MB/s, file downloaded in 1 min.
October 13, 2025 at 4:02 PM
The year is 2025: can't find a single browser that uses multiple connections to download a file from a direct link. Not even 1 MB/s on the browser, about 30 mins to finish the download (2.5 GB file). Send it to `axel` with 30 connections: 44 MB/s, file downloaded in 1 min.
Aaaand he also wanted to lock RubyCentral out of their own AWS account? Totally trustworthy person to host your gems, guys.
October 9, 2025 at 8:45 PM
Aaaand he also wanted to lock RubyCentral out of their own AWS account? Totally trustworthy person to host your gems, guys.
Ohhh, so this guy that started the alternative RubyGems server wanted to have access to HTTP logs to sell behavior data to 3rd parties???? WTF.
October 9, 2025 at 8:42 PM
Ohhh, so this guy that started the alternative RubyGems server wanted to have access to HTTP logs to sell behavior data to 3rd parties???? WTF.
Holy guacamole, what is this????
October 9, 2025 at 3:25 PM
Holy guacamole, what is this????
I'm sorry, but there is no reason to pay any of the Anthropic plans. Particularly true if you are on the 100 or 200 USDs plan. Just get ChatGPT Pro. GPT-5 (Codex) is incredible and the limits are good even with high thinking effort. Claude Code 2.0 is a nice TUI though.
October 3, 2025 at 2:06 PM
I'm sorry, but there is no reason to pay any of the Anthropic plans. Particularly true if you are on the 100 or 200 USDs plan. Just get ChatGPT Pro. GPT-5 (Codex) is incredible and the limits are good even with high thinking effort. Claude Code 2.0 is a nice TUI though.
Anthropic’s new limits kicked in and Claude Code users are having a meltdown all over X and Reddit. Looks like in 2 days people got up to 90% of the limit already. Some got there in a single day. I guess there will be a massive migration to OpenAI thanks to this and GPT-5. 😂😂😂
October 1, 2025 at 10:32 PM
Anthropic’s new limits kicked in and Claude Code users are having a meltdown all over X and Reddit. Looks like in 2 days people got up to 90% of the limit already. Some got there in a single day. I guess there will be a massive migration to OpenAI thanks to this and GPT-5. 😂😂😂
After seeing all the issues and outright broken design in the latest OS releases by Apple, does anyone have faith that they have an LLM app that is actually on-par or better than ChatGPT's? They could just be quiet and release something good. They prepare the jokes themselves.
September 26, 2025 at 10:27 PM
After seeing all the issues and outright broken design in the latest OS releases by Apple, does anyone have faith that they have an LLM app that is actually on-par or better than ChatGPT's? They could just be quiet and release something good. They prepare the jokes themselves.
New iOS and macOS 26 *stable* releases full of issues, from bad performance to poorly designed interfaces (even including text that is not aligned). What a stain in the reputation of a company that was known for well designed and high quality software. 🤮
September 19, 2025 at 12:30 PM
New iOS and macOS 26 *stable* releases full of issues, from bad performance to poorly designed interfaces (even including text that is not aligned). What a stain in the reputation of a company that was known for well designed and high quality software. 🤮
Even my parents internet connection in Brazil has IPv6 connectivity, but not Movistar/O2 still lacks it in Spain. 😳
September 19, 2025 at 12:18 AM
Even my parents internet connection in Brazil has IPv6 connectivity, but not Movistar/O2 still lacks it in Spain. 😳
Garmin is so out of touch with their UX. The LiveTrack view always defaults to showing distance in miles. If I recall correctly, I can count on one hand the amount of countries using miles for distance. Why isn’t the default distance unit in meters (like in over 100 countries)?
September 2, 2025 at 11:31 AM
Garmin is so out of touch with their UX. The LiveTrack view always defaults to showing distance in miles. If I recall correctly, I can count on one hand the amount of countries using miles for distance. Why isn’t the default distance unit in meters (like in over 100 countries)?
Awesome to see @homebrew getting concurrent download support. You can try it out with the env var `HOMEBREW_DOWNLOAD_CONCURRENCY`. Would be lovely if we also got concurrency for installation/upgrades of bottles that are independent of each other in the future.
September 2, 2025 at 8:21 AM
Awesome to see @homebrew getting concurrent download support. You can try it out with the env var `HOMEBREW_DOWNLOAD_CONCURRENCY`. Would be lovely if we also got concurrency for installation/upgrades of bottles that are independent of each other in the future.
There's isn't a better article out there to understand why LLMs as of today are a bubble and how much it'll hurt when it pops:
The Hater's Guide To The AI Bubble
Hey! Before we go any further — if you want to support my work, please sign up for the premium version of Where’s Your Ed At, it’s a $7-a-month (or $70-a-year) paid product where every week you get a premium newsletter, all while supporting my free work too.
Also,
www.wheresyoured.at
August 12, 2025 at 8:39 AM
There's isn't a better article out there to understand why LLMs as of today are a bubble and how much it'll hurt when it pops:
A few years ago, people used to joke about AMD CPUs and anyone buying them over Intel's. Intel had the upper ground for many years. Today it's the opposite: Intel seems to be a high risk of completely falling apart (if the US gov doesn't rescue them) and AMD dominates CPUs.
August 11, 2025 at 5:31 PM
A few years ago, people used to joke about AMD CPUs and anyone buying them over Intel's. Intel had the upper ground for many years. Today it's the opposite: Intel seems to be a high risk of completely falling apart (if the US gov doesn't rescue them) and AMD dominates CPUs.
What's wrong with all the VS Code extensions using helm-ls? In my Neovim setup that uses Mason, it can instantaneously jump to the definition of anything. In VS Code using the helm-ls extension it just can't. :/
August 8, 2025 at 9:49 AM
What's wrong with all the VS Code extensions using helm-ls? In my Neovim setup that uses Mason, it can instantaneously jump to the definition of anything. In VS Code using the helm-ls extension it just can't. :/
GPT-5 bingo card: "PhD level intelligence". Literally everything is "PhD level intelligence" nowadays, no?
August 7, 2025 at 5:02 PM
GPT-5 bingo card: "PhD level intelligence". Literally everything is "PhD level intelligence" nowadays, no?
Wow, @zed.dev is getting neater by the day. Beautiful, smooth, useful. It's a delight every time I use it for a bit to check their progress.
August 6, 2025 at 10:04 AM
Wow, @zed.dev is getting neater by the day. Beautiful, smooth, useful. It's a delight every time I use it for a bit to check their progress.
"Introducing our (...), a new AI-powered platform that lets you create apps (...)" is the new Javascript framework: every day there's a new one. Nobody uses them. Not sure how they survive. Everyone wants to build one though.
August 5, 2025 at 4:40 PM
"Introducing our (...), a new AI-powered platform that lets you create apps (...)" is the new Javascript framework: every day there's a new one. Nobody uses them. Not sure how they survive. Everyone wants to build one though.
It's been almost a full year since the announcement of the Elixir LS project. Is it ready for daily usage? Which LS should I use for Elixir/Phoenix nowadays? 😬
July 31, 2025 at 1:55 PM
It's been almost a full year since the announcement of the Elixir LS project. Is it ready for daily usage? Which LS should I use for Elixir/Phoenix nowadays? 😬
There should be a benchmark to measure how good an LLM based on what I'm calling the "You're absolutely right" Index (YARI). Claude just literally told me this 4 times in a row and still didn't fix its mistake that I outlined pretty clearly.
July 29, 2025 at 3:17 PM
There should be a benchmark to measure how good an LLM based on what I'm calling the "You're absolutely right" Index (YARI). Claude just literally told me this 4 times in a row and still didn't fix its mistake that I outlined pretty clearly.
The Claude Max plan fiesta is over. Weekly limits are here! 😅
July 28, 2025 at 6:49 PM
The Claude Max plan fiesta is over. Weekly limits are here! 😅