Pretty sure I saw some comments saying it was too inconvenient. Frictionless experience.. Convenience will likely win out despite any insanity. It's like gravity. I can't even pretend to be above this. Even if one doesn't use these things to write code they are very useful in "read only mode" (here's to hoping that's more than a strongly worded system prompt) for greping code, researching what x does. How to do x. What do you think the intention of x was. Look through the git blame history blah blah. And here I am like that cop in Demolition Man 1993 asking a handheld computer for advice on how to arrest someone. We're living in a sci-fi future already. Question is how dystopian does this "progress" take us. Everyone using llms to off load any form of cognitive function? Can't talk to someone without it being as common place as checking your phone? Imagine if something like Neuralink works and becomes ubiquitous as phones. It's fun to think of all the ways Dystopian sci-fi was and might soon me right
I was replaying Cyberpunk 2077 and trying to think of all the ways one might have dialed up the dystopia to 11 (beyond what the game does). And pervasive AI slop was never on my radar. Kinda reminds me of the foreword in Neuromancer bringing attention to the fact the book was written before cellphones became popular. It's already fucking with my mind. I recently watched Frankenstein 2025 and 100% thought gen ai had a role in the CGI only to find out the director hates it so much he rather die than use it. I've been noticing little things in old movies and anime where I thought to myself (if I didn't know this was made before gen ai, I would have thought this was generated for sure). One example (https://www.youtube.com/watch?v=pGSNhVQFbOc&t=412) cityscape background in this a outro scene with buildings built on top of buildings gave me ai vibes (really the only thing in this whole anime), yet this came out ~1990. So I can already recognize a paranoia / bias in myself and really can't reliably tell what's real.. Probably also other people have this and why some non-zero number of people always thinks every blog post that comes out was written by gen ai.
I had the same experience, watching a nature documentary on a streaming service recently. It was... not so good, at least at the beginning. I was wondering if this was a pilot for AI generated content on this streaming service.
Actually, it came out in 2015 and was just low budget.
>I do hate that if you publish anything online these days, someone will accuse you of having used AI to write it.
Noticing this too. Sabine said something a while ago in one of her videos that stuck with me [0]. about people expecting proof of suffering by next year. She was talk submitting an essay, but it might as well be anything ai could have done.
I worked for a company that.. Used msql sever a lot and we would run into a heisenbug every few months that would crash our self hosted msql server cluster or it would become unresponsive. I'm not a database person so I'm probably butchering the description here. From our POV progress would stop and require manual intervention (on call). Back and forth went on with MS and our DBAs for YEARS pouring over logs or whatever they do.. Honestly never thought it would be fixed. Then one time it happened and we caught all the data going into the commit and realized it would 100% reproduce the crash. Only if we restored the database to a specific state and with this specific commit it would crash MS SQL Server. NDAs were signed and I took machete to our code base to create a minimal repro binary that could deserialize our data store and commit / crash MS SQL sever. Made a nice powershell script to wrap it and repro the issue fast and guess what? Within a month they fixed it. Was never clear on what exactly the problem was on their end.. I got buffer overflow vibes, but that's a guess.
I once ran into a bug where our server code would crash only on a specific version of the Linux Kernel under a specific version of the OpenJDK that our client had. At least it would crash at startup but it was some good 2 weeks of troubleshooting because we couldn't change the target environment we were deploying on.
At least it crashed at startup, if it was random it would have been hell.
small feedback for the readme: I think demo should be above the explanation on the differences in the port. Perhaps it doesn't matter as this might just be a thing people find after they know what they're looking for. I didn't really get what it is until seeing the demo and still had to lookup what a webring is on Wikipedia because I was curious why it's called that. Anyway cool looking project. I might even use it if I ever stop procrastinating and blog again.
BTW your blog is probably the most pleasant looking website I've seen in a month at least to my eyes on mobile.
Dunno if you're right, but I'd like to point out that I've been reading comments like these about every model since GPT 3. It's just starting to seem more likely to me to be a cognitive bias than not.
Why would you assume cognitive bias? Any evidence? These things are indeed very expensive to run, and are often run at a loss. Wouldn't quantization or other tuning be just as reasonable of an answer as cognitive bias? It's not like we are talking about reptilian aliens running the whitehouse.
I'm just pointing out a personal observation. Completely anecdotal. FWIW, I don't strongly believe this. I have at least noticed a selection bias (maybe) in myself too as recently as yesterday after GPT 5.1 was released. I asked codex to do a simple change (less than 50LOC) and it made a unrelated change, an early return statement, breaking a very simple state machine that goes from waiting -> evaluate -> done.
However, I have to remind myself how often LLMs make dumb mistakes despite often seeming impressive.
I haven’t noticed the effect of things getting worse after a release but definitely 2.5’s abilities got worse. Or perhaps they optimized for something else? But I haven’t noticed the usual “things got worse after release!” Except for when sonnet had a bug for a month and gpt5’s autorouter broke.
Sometimes it is just bias but the 2.5 pro had benchmarks showing the degradation (plus they changed the name every time so it was obviously a different ckpt or model).
reply