Nick Blow (@NickBlow) header
Nick Blow (@NickBlow) profile picture

Cofounder @ https://teapawt.net/t.co/lDCAkm19bI. Prev, @iterate, CTO @Wakelet. Dad to 2 daughters. No relation to Jonathan Blow. Implicit IMO.

On my laptop
Joined January 2024
Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)Photo by Nick Blow (@NickBlow)
Wrote about AI detection mostly to get my own frustration about the topic out... nickblow.tech/posts/lets-tal… We need more, and better AI detectors – not just Pangram. For places where the words are the work (such as technical writing or publishing), we should absolutely be demanding proof of human effort. I don't feel like it needs to be 100% human, but it should still be overwhelmingly so, each word picked with intention. And some of the loudest voices against this are those who have the most to lose if it turns out they're just laundering Claude tokens.
3
81
It’s so frustrating that the idea of AI detectors being unreliable has finally reached people just for AI detectors like pangram to be actually useful… and on top, the average user isn’t familiar with Claudeslop so they can’t even see it when it’s right in front of their eyes.
1
2
143
Nick Blow retweeted
[read about what’s happening with AI] ME: Wow we’re actually legit fucked this is so bad. [go outside] ME: Wait those people are all acting like nothing unusual is happening. Maybe I was just in a paranoid bubble? Yeah it must be that. Phew. [read more about what’s happening with AI] ME: Wait no we ARE fucked. Shit. It’s all those people outside who are in a delusional bubble. [go outside] ME: Wait...
369
1,007
182
14,585
720,477
Thermonuclear take, better models make it harder to have good code. Opus5.5 is phenomenal, it’s the best model out there. But it still does odd things, still duplicates a lot, still makes mistakes. The biggest problem, imo, is that because let’s say 75% of what it does is *fine*, it’s a lot harder to stay mentally engaged when reviewing it. If there’s a problem every few lines or major obvious flaws, then it’s a lot easier to focus because you’re thinking about how to do it better. Now that the PRs are mostly fine, I’m constantly fighting the urge to just merge without major revision, especially for “less important” bits of the codebase. And it’s not quite good enough to do that safely.
13
1
38
8,874
Made a steak pie today, pretty good 👍
1
78
kinda mad how cloud run is such a good product AND no one has done any serious competition to it for years AND google cloud is so bad generally otherwise
1
97
On the one hand I’m glad that I’ve been focusing heavily on using intelligence without increasing tokens by creating reusable code snippets; on the other hand it’s just bad overall that data centres look like they’re going to be heavily restricted. The only silver lining is it does seem like the anti data-centre vitriol is slightly dying down, so perhaps there are solutions… Space data centres are likely inevitable, and it’s sad that we need to solve incredibly complex engineering problems because it’s easier than dealing with bureaucracy. Potentially likely is some kind of legal wrangling as to what exactly constitutes a data centre, and then creative solutions, perhaps decentralised compute by paying people for idle GPUs in their PCs(which, granted, is an enormous software engineering challenge). Other countries are an option, but I’m not sure which would even be viable, as they have similar / worse levels of bureaucracy… TL;DR tokens are likely gonna get more expensive and that sucks
3
153
I find it significant that the most AI-pilled people I know use AI for basically everything…except for writing It’s remarkably consistent. Why? Because to write is to think. And outsourcing your critical thinking to AI is intellectual suicide
208
226
62
4,432
255,494
Occasionally I wonder what the first game I played with AI code in it. Or if I have played one yet? My gut feel is that the resurgence in “classic” game revivals is heavily influenced by AI being able to navigate decades of spaghetti 🍝. I’d assume every patch post let’s say 2025 has at least some AI code in it, but in terms of base game it’s a little harder to say as dev cycles can be long.
1
3
205
I do wonder what’s going to happen with public data with Reddit going logged-in only. There’s this weird tension between agents and sites not having easily usable, pay-per-use APIs. But again, the RoI for those might not be worth it or obvious for the big providers as agents are still niche and they also can’t view ads…
1
123
On a sales call today, and turns out they had used the product I built as CTO of a previous startup (Wakelet), so that's pretty cool! Even better, they had good things to say about it 😅
1
3
141
Few more observations on working with AI-enabled team: 1. One of the things that separates junior devs from senior is the ability to notice that a design is going wrong and trying another direction early. Juniors will push in a hopeless direction for way too long. AI makes it worse. Juniors feel like AI validated the approach, avoid talking to others (most hate it) and push in the weird direction for too long. 2. When it comes to code produced by seniors, reviewing code is rarely useful. Reviewing the intent of PRs and whether I agree with the intent is important. Are we fixing the correct problem? in the correct part of the code? In a way that is friendly to users, operators, the company? 3. Work now feels... lonely. Code reviews (and sometimes even design) are mostly comments from AI, fixed by AI. Engineers always saw any conversation as waste of time, when they could just produce more code. Now they can produce even more code, and it seems like a conversation with me adds less value than talking with AI, and perhaps even less pleasant. It isn't about being social at work. They are happy to talk about movies, news, etc. But I find conversations with engineers about engineering topics to be energizing, inspiring and motivating. And while I understand it isn't the job of my team to energize and inspire me, I deeply miss the interactions we used to have.
37
55
7
663
30,915
No fucking way ITS 4 AGAIN DIFFERENT BUG
3
152
Quick thoughts from a few weeks of moderate production usage of decision models: Jev is the best bang for your buck, still. It’s remarkably good and remarkably cheap. OpenAI decision API is legit, cheaper than Luna while still performing very well. In my tests, this was the best decision model. More expensive than Jev, and slightly high etc latency. Clef is sadly not great, all my evals had clef get false positive/negative on 30-40% of the options, whereas Jev and OpenAI were very close to 100% accurate. And it’s not cheap enough to bother ripping out Jev. Maybe if you are entirely locked into Cloudflare workers AI, but I suspect the OpenAI model will come there soon too…
1
139
Don't forget: you're no longer working with humans
65
148
37
2,125
121,821
Did anyone zero day the switch2?
1
4
143
Oh shit, this is sad. The first video game I ever completed was Knights of the Old Republic. I’d started many others but none had hooked me. Mass Effect 1&2 are also genuinely two of the best games ever. They’ve had a series of duds, but damn. I guess we’re never gonna see ME4.
BREAKING: Electronic Arts is cutting 400 jobs and shutting down BioWare, the legendary development studio behind Dragon Age and Mass Effect. bloomberg.com/news/articles/…
2
1
133
It's really rather lame that both Apple Music and Spotify APIs explicitly ban synchronization to other software (e.g. games or visualizers)
1
4
283
App store review is such a load of BS. I made an app for my daughter to learn german vocab. I'm currently of round 3 of review back-and-forth because my walkthrough video is not good enough. I just asked claude to make it a PWA and it was up and online 20 min later...
8
1
1
28
1,848