←
PatrickUllrich1,377 posts
PatrickUllrich avatar

PatrickUllrich

@ItsPatrickU

🇨🇦 🇩🇪

◷ Joined October 2019▧ 60 Photos and videos
𝚓𝚒𝚖𝚖𝚢 avatar
⟳ PatrickUllrich retweeted
𝚓𝚒𝚖𝚖𝚢@202accepted·Oct 24

a lot of guys in SF are about to start dating their macbooks

1:52
jack@jackOct 23

We believe voice is the most natural way to interact with intelligence, so we created Dessa, the most authentic full-duplex speech model ever built. Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence. block.xyz/dessa

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 24

Engineers are doing 2000x times as many PRs but why does it feel like every company still ships just as much value as before?

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 23

Architecture still matters, because it makes verification cheaper and raises throughput, because it lowers the impact and complexity of PRs. Meaning less PRs require human in the loop because of it.

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 18

Should make this clear. I don’t think there’s anything wrong with it, just that Silicon Valley rarely represents the rest of software development. Jfc I worked with “senior devs” (especially consultants) where I was happy they could open an IDE let alone ship usable code. The other huge difference is, while there’s truly some incredible engineers in SF, there’s just as many entrepreneurs. Again, nothing wrong with that, just keep the context in mind when evaluating people’s tech takes / opinions.

I think this is true, but also on many things they have been pioneers.

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 17

Not to pick on Thorsten here, but this (and much of the current discourse) is a great reminder that SF companies building dev tooling, where their own team is the customer, aren't representative of the rest of the industry. The context these teams and people work in is so inherently different that it's tricky to have a fruitful conversation. 160 characters aren't enough to agree on a framing first.

Sure sounds like a lot of people don't trust their colleagues to work like this was their own project.

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 15

MCP that helps your bot to find clothes that: - are available in your size - ship to your country - fit in your budget Actually look like in the picture. UI to illustrate the service

0:27

Why can ChatGPT/Claude generate outfits, but then can’t find me the clothes online for me to check out? 🙃

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 12

Been a professional @AirCanada hater for a while. While I am still salty of their questionable business practices during COVID, I gotta admit they really stepped it up over the last while. The app is easily the best airline app there currently is. And service seems like it has improved as well. 🫡

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 11

Missing from the whole AI conversation: what does the career path for junior engineers look like now? For senior devs this is heaven. We finally get to build all the things we had in our heads. But we only had those in our heads because we ran into the problems before. Juniors never experienced that part. I've watched a lot of people's growth stall after 3-6 months, and their output with the models hit a ceiling. So how does a junior become a senior without coding? Reminds me of math in grades 1-4. Calculators always existed, but we had to do by hand for a reason.🤷‍♂️

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Oct 11

Why can ChatGPT/Claude generate outfits, but then can’t find me the clothes online for me to check out? 🙃

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Sep 14

And another s/o to @t3dotcodes for being open source (and easy to fork locally). Every one of these experiments is basically a prompt away, and the product is already good enough that I can use it in my actual day-to-day work while trying these ideas as a POC. That makes it a pretty fun place to test out what a better PR review workflow could look like.

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Sep 14

5. Funny enough, the more I look for ways to understand a change beyond just reading the diff - given I don’t fully trust an AI agent to do the review for me yet - the more I keep finding ways to give the agent better context. And at some point, that might actually make it better at reviewing the PR than I am. 😅

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Sep 14

Honorable mentions: DDD, C4, dependency graphs, etc. DDD feels useful as a lens for asking “did the domain concept actually make it into the code?” C4 is tougher. It’s great for understanding a system, but it’s intentionally high-level/static. For the average small PR, I’m not sure a freshly generated C4 diagram tells me much I didn’t already know. (See #1)

PatrickUllrich avatar
PatrickUllrich@ItsPatrickU·Sep 14

3. Abstraction pressure. Excited about this one. While you shouldn't be building for future features, I think running your Abstraction through a few hypothetical upcoming changes is an interesting way to qualify whether your code is good or not. Another skill add.

Post attachment