@mlevison@agilealliance.social
Post #414970
2026-02-04 20:55 UTC
Replies (19)
-
@mlevison@agilealliance.social 2026-02-04 20:55
The bottleneck was never writing code. It's understanding what to build. If you're using AI coding tools, focus on: • Smaller features (if it's 1000 lines, it's too big to review) • Clear acceptance criteria before you prompt • Tests first, AI-generated code second • Security audits (AI can't do this) More code isn't the goal. Solving real problems is.
-
@airwhale@beige.party 2026-02-04 21:53
@mlevison Amen to that, brother…
-
@LunaticDancer@mastodon.gamedev.place 2026-02-05 07:44
@mlevison so true, it astounds me people didn't see this coming from the start. Also there's the cognitive deterioration double whammy.
-
@oschonrock@mastodon.social 2026-02-05 07:45
@mlevison That's my experience too. Dave Farley's. MSE channe, which I usually respect,l recently claimed the opposite though, based on a study they took part in.
-
@piper@lgbtqia.space 2026-02-05 07:49
@mlevison Goodhart's law in action. Actually, I'm not sure it even is an example of Goodhart's law. Raw quantity of code output would never have correlated strongly with quality. What do they think they're doing‽
-
@crackhappy@cyberpunk.lol 2026-02-05 08:07
@mlevison I use LLMs to help me with basic code writing tasks, generating the structural frameworks, saving me a lot of typing time. However, I never rely on that code out of the box, I always review it thoroughly and often just snip and prune. I would never attempt to give an LLM a complicated set of instructions, it's going to fail every time.
-
@thirstybear@agilodon.social 2026-02-05 08:27
@mlevison I hear you. I am seeing it first hand. And I am being hammered for pointing this out and encouraging caution, while others are being rewarded for shipping 💩 “because AI”. Just call me ‘Cassandra’ 🤷♂️
-
@H4Heights@mstdn.social 2026-02-05 08:53
@mlevison l suspect the line about the volume of code increase is conservative.
-
@xabaras@androiddev.social 2026-02-05 08:55
@ivanmorgillo What's your experience about this?
-
@enigma@norden.social 2026-02-05 09:42
@mlevison Question : would have moonlanding been also a success if AI would have had control ? 😄
-
@deborahh@cosocial.ca 2026-02-05 11:13
@mlevison @Em0nM4stodon → 30-41% increase in technical debt 😱
-
@Minnesota411988@mastodon.social 2026-02-05 14:57
@mlevison Defining “productivity” as building the wrong thing faster has been the usual definition for at least my lifetime. Not just for code.
-
@rhold@norden.social 2026-02-05 15:01
@mlevison can you post some links to those research papers? In case you did not know: #acm's digital library is not open and free to use for everybody so most peer reviewed articles published in the us should be accessible. https://acm.org/dl
-
@minego@pdx.social 2026-02-05 16:45
@mlevison The silver lining here (for developers who aren't using AI, despite demands from employers) is that we are easily able to keep up with developers who are. My boss saw speed gains from some developers, and asked us all to. I kept going as normal, and after a bit of time for things to settle, I'm still faster. I like to think that more and more people are realizing that the emperor has no clothes.
-
@viktor_linsel@linke.social 2026-02-05 11:47
@mlevison A problem of capitalism: it's always about *how much* you produce and consume, and rarely about *what* you produce. It seems that problem has now reached coding too.
-
@yon@sakurajima.moe 2026-02-05 18:51
@mlevison My anecdotal guess is that people are operating above their own abilities. If I understand a code base and its architecture, dependencies, business logic needs, etc. Then I can ask an AI to make a targeted change and verify it. It’s basically avoiding some typing. But if I don’t understand those things I don’t actually know how to ask or how to verify it. Even worse if I wouldn’t be able to write it in the first place, how can I ask for it and verify it? Basically, how can I successful operate above my own knowledge and ability? It’s like asking for a piece of text written in a language I don’t speak. How can I have a clue what it says? I think there are so uses for assisted development. I’m less sure an LLM is the way to achieve that. Worst part, people aren’t learning anything from all this. There’s no skill and knowledge being built:( Imho, ymmv, etc. Also those are some numbers, yikes!
-
@jose@mastodon.gamedev.place 2026-02-05 19:03
@mlevison@agilealliance.social a good friend recently started experimenting with claude, and he gave me the most blunt statement about the output: it's instant legacy code. That's it. It really resonated with me, for so many reasons.
-
@steltenpower@social.edu.nl 2026-02-05 19:55
@mlevison@agilealliance.social profucktivity?
-
@elly@donotsta.re 2026-02-05 22:08
@mlevison@agilealliance.social Just plain 'ol snakeoil If you know to code (especially in area of interest/what's needed for the project)... why would you decide to waste your time arguing with slop machine? Everyone who tried LLMs tells me "it saves me time", then I end up wasting my time reviewing absolutely pointless/useless/broken patches or nonsense conversations with overclocked text predictor (if it's a client, I don't bother otherwise). On top of that people who use LLM experience loss of cognitive abilities, don't learn anything new, don't understand code that slop presents them with and exhibit narcissistic behavior (especially when confronted about slop they're trying to merge). There are *zero* valid use-cases for Generative AI*. *Using "AI" as accessibility features like live transcriptions doesn't count. **Decades-old pattern recognition/computer vision technology doesn't count as modern "AI" hype and has valid use-cases.