Adam McNeilly

@adammc.bsky.social

Android GDE | Airbnb Find me on Twitch: https://twitch.tv/adammc

At Tuesday's meetup, I was asked a question about agentic coding and I responded "I'm the wrong person to ask." When they wanted to know why, I said it's because I don't use it. They were shocked at first, but then admitted "I feel like it's making me dumber." I heard this from at least two people.

I am exasperated with people accusing programmers of "gatekeeping" programming. We were practically dragging people through the gate. Do you know why a mindless LLM can tell you so much great programming information? It's because we've been sharing that information for free online for decades.

I just called to schedule a service appointment for my car, and they used an automated system which is FINE but if you're going to do that, at least say it up front. Don't say "I'm Alex" and talk to me like a human being when the whole interaction is clearly automated and rigid.

If you're a manager at any company mandating LLM use, I'd recommend you try to ask your directs how they *really* feel about LLM adoption, promising to shield them from retaliation. Because it can be *very* costly for one's career to criticize LLMs right now. Empathize with your workers.

Honza Dvorsky@czechboy0.dev · 2mo ago

It's sad how the top-down LLM mandates at companies created an atmosphere of fear for and retaliation against anyone even mildly skeptical of LLMs. I get so many engineers reaching out to me privately (please continue doing so, these messages make my day), afraid to speak up publicly. 😕

As a science fiction writer I really fucking resent tech bros for ruining the phrase "artificial intelligence." 80 years of it being all about cool androids and interesting computers destroyed by five years of "Spicy autocorrect making everything shittier for money." NEVER FORGIVE.

> Even if the code is written for you there's still a huge cognitive load in tracking all of that output. I think AI ultimately increases cognitive load. Instead of an implementation I can already do, I wait for AI to write and then read from scratch. Previously, we understood it while writing it.

Ryan@projectpixel.net · 3mo ago

Doesn't this just mean you're going to be expected to churn out more and more work in less and less time? It's not like you'll get the benefit of getting it done quickly the having a break. Even if the code is written for you there's still a huge cognitive load in tracking all of that output.

I turned off the keynote. I feel like I'm going certifiably insane. They admit that people can detect high quality AI only 25% of the time, and so their solution is to watermark AI content, with a watermark invisible to the naked eye? That solves absolutely nothing. I'm disgusted, honestly.

Setup my walking pad at my desk today so I could walk while watching Google IO, and then got the cursed exercise idea of "what if I increase the speed every time Sundar says AI" 💀

Last Friday, we announced a ✨ new default structure ✨ for Kotlin Multiplatform projects. (Nothing like a late Friday afternoon release!) In case you missed that, here’s a quick recap of what’s new! 🧵👇