Behzad retweeted
⚠️⚠️⚠️ WARNING: IF YOU HAVENT DONE SO ALREADY, BREAK THE SCREEN TO YOUR CELL PHONE NOW !!!
BREAK THE SCREEN TO YOUR CELL PHONE NOW.
this is crazy my session cache hit rate is almost 100
there is a psychosis going on with swe that telling the claude to make you an app is a long term skill set that is going to pay hundreds of thousands of dollars a year. some people will make bank before economy readjusts to software being this easy to make but it most probably won't be you
You’ve just ordered Pizza Hut and 2L of Mountain Dew. You’ve loaded WoW on your crappy PC. Somehow, your paladin has reached level 34, and you’re about to group up for a quest in Stranglethorn Vale. Your parents don’t mind you staying up all night. A perfect summer night. You are 35 years old. The year is 2026.
Behzad retweeted
Fork your dependencies, trim them to only your use case, never update unless it breaks for your users. I’ve been vocal about this for 10+ years. I’ve always said that updating is way riskier than latent bugs (which can be tracked and CVEs monitored).
If you are updating a dependency, it’s on you to analyze every single commit in the full transitive set of dependencies. If you dont see anything compelling, dont update!
I remember at HashiCorp once in awhile an engineer would try to update a dep or replace a DIY lib with an external one and id always ask “show me the commit we need.” Dont update for the sake of it.
Feeling pretty swell about this mentality with all the supply chain attacks happening.
for me it was mostly about gamifying work for myself so it takes less energy and is more fun. neovim is great for that
I think the thing people don't understand about how I used Neovim and what I liked about it (even though it's what I repeatedly said many many times) is this:
I felt like it was an effective tool for solving real problems and delivering value for people on a consistent basis for a long time.
Some of that value was derived from my own personal feelings and some of it was derived from things that I thought I could do faster. Some of them were just about enjoying my job, like having a ping pong table at work.
I will continue to use whatever tools I think allow me to do the best job that I can over an extended period of time, deliver value for people, stay employed, and keep putting food on my family's table.
Of course right now that mostly means answering emails for the Primeagen, but some day that might mean coding.
GLM-Legal uncensored:
Introducing GLM-Legal
Built on GLM-5.3-Flash, GLM-Legal matches frontier models on mainstream legal benchmarks.
Designed for real legal workflows: contract review, evidence analysis, compliance, and more.
Tech Blog: z.ai/blog/glm-legal
Apply for API Access: z.ai/consultation/legal
Try AutoClaw: autoclaw.z.ai
is jetbrains the biggest fumble in the entire dev tooling history?
always hostile to open standards like lsp, way too late to the ai coding space while having a head start
it's crazy
mostly made me feel dread. used to be able to go from question to typing out some solution in seconds, way slower now.
what's gonna happen when we forget how to code well and we collectively provide the training data to the labs for the parts ai is not good at right now by non stop interaction with ai?
"ai is good at coding but not good at x"
you should get ready to either keep moving the goalposts indefinitely or accept that it will surpass any specific criterion you set for "x" and what that means for you
i'm kinda surprised @t3dotcodes hasn't added a orb like cloud service yet. it would fit the app very well
one hard part of making products with ai is avoiding every single one becoming an everything app. any feature I need can be added to my app with ai pretty fast but it would ruin the product in the long run























