Aftershock Festival Lineup
Wow. This lineup is wild. #AftershockFestival
aftershockfestival.com
05.03.2026 04:03
👍 1
🔁 0
💬 1
📌 0
Aftershock Festival Lineup
Wow. This lineup is wild. #AftershockFestival
aftershockfestival.com
Do I trust it to do this for me in an automated pipeline? Definitely not.
Will I use this for persisting hand drawn diagrams in the future? Probably yes.
Gemini impressed me today.
I uploaded a picture of my whiteboard and asked it to generate xml for a draw.io diagram based on the image. First in cleaned up the image into something like clipart and then it generated the xml.
Uploading the xml into draw.io worked first time and was 95% correct.
Matched
Personally, I hate having to lie to compete.
Sometimes I opt for lies of omission to assuage my conscience.
But we live in a world where many people lie to get ahead, or even just for fun. You’re probably working with people like that already.
Are you that person?
Was it worth it?
Interviewing is a sales meeting.
You’re selling yourself.
Some people lie to sell. It can increase their chances of closing, if they get away with it.
Some people despise lying. These people avoid the risk of being caught but increase the risk of failing to sell.
www.reddit.com/r/jobsearchh...
This is a little more optimistic about AI than I feel, but I like the emphasis of skepticism as a virtue. Optimists often mistake skepticism as pessimism. But we could all use a healthy dose of skepticism to prove out our ideas and help validate our decisions.
This post by @swardley.bsky.social expresses some of my fear of using AI to code.
It doesn’t just generate tech debt. It generates a debt of understanding. It doesn’t understand the dependencies and consequences. It’s not using reasoning to make good decisions.
medium.com/mapai/on-the...
This is how I feel about most coding agents today.
To be fair, it kinda feels like even with exclusively human authors, most coding projects eventually become an unmaintainable mess.
Agents just get there faster, sometimes before it ever actually works.
youtu.be/2_b8HM-OfMU
Fair. I do like me some live jazz and beach sunsets.
Seems like a supply and demand problem.
Vendors used to pay money to be more secure.
Now everything is insecure and finding new vulnerabilities is trivial.
The industry doesn’t need more bug finders. We need more bug fixers, and companies willing to pay for more security.
Noisy Neighbor Detection.
Bookmarking for future exploration.
github.com/softcane/Kub...
Reviewing inputs into abstraction layers is pointless if you don’t trust the abstraction layers.
GitOps of infra and platform config is full of lazy half abstraction layers that don’t have schema or tests… multiple layers of them.
Why do we do this to ourselves?
I’m totally down for fantasy escapism!
Reddit is overrun by bots and assholes.
Facebook is selling your eyeballs to propagandists.
X is full of nazi fascist pedophiles.
Instagram is full of fake influencers.
TikTok is full of misinformation.
And Bluesky is depressingly real.
Where do you go for good vibes?
Depression on demand!
What do I need social media for when I can just ask ChatGPT to simulate social media for me?
But why can’t they just take my word for it? That would be so much faster and easier.
Yes, I’ve heard the standard advice:
- Pick your battles
- Collaborate instead of correcting
- Make them think being wrong is their idea by asking questions
How do you tell people they’re wrong when telling people they’re wrong gets you put in the doghouse?
Do you just let them be wrong when you know it’s going to be a problem for you and your team later?
I have discovered the easy way to detect a human: ignorance.
“It doesn’t really matter if it was on purpose or not. If it was a deliberate acting choice, then it’s a brilliant one. If it came purely out of him trying to steer into the skid, then it’s still brilliant.”
I wrote a lot of code at Google, but I think this StatusWatcher might have been my favorite.
github.com/kubernetes-s...
