piazz

6 ms·

1.
by piazz 3d ago
I’ve been hacking on a single Anki extension for… the last three years now: https://smart-notes.xyz The idea is to make an extension that allows you to flexibly fill out fields on your cards with AI generated text, speech, and i
2.
by piazz 23d ago
Can we CRISPR it into humans? I volunteer. I’ve always wanted to improve my protein folding accuracy.
3.
by piazz 25d ago
I'm using a slimmed down version of this skill: https://github.com/steipete/agent-scripts/blob/main/skills/c... The important bit I found is to explicitly remind Claude that Sol 5.6 is a very
4.
by piazz 26d ago
Summer 2026 meta (well, mine): Sol is for routine work, Opus for frontend/design, and Fable for more complex / ambiguous / architecture work. Fable works extremely well to drive Sol as a subagent. Fable is the only one you ca
5.
by piazz 1mo ago
> The implications are clear: focus on the human-skills part of the job. The implications are not clear. They are not clear for the security people, for the SWE people, for anybody in knowledge work whose jobs are impacted. I wish people
6.
by piazz 2mo ago
Great writeup. The excessive function thing has always driven me crazy; I guard against this explicitly in Claude.md. I have found that models are generally poor at managing refactors / complexity while also implementing new features
7.
by piazz 2mo ago
> This is where OpenAI has an advantage over Anthropic. While its models are trailing Anthropic's in recent months, its investments in product, consumer experience, site publishing, voice, and hardware are all directions that have c
8.
by piazz 2mo ago
This one observation changes everything.
9.
by piazz 3mo ago
PSA this is satire ;) (if you have to say it, that’s how you know it’s good)
10.
by piazz 3mo ago
The term that I saw once, and now constantly land back on, is “meaning shaped” At a quick glance, it looks like a thing that should contain meaning and substance. At any closer inspection, it falls apart completely.
11.
by piazz 4mo ago
“Honestly” / “the honest answer is” are huge LLM tells. Spend enough time arguing with Claude and hearing that combination of words starts making you wince / twitch uncontrollably. That said I enjoyed the article!
12.
by piazz 4mo ago
100% agree. Took me ages of working with the agents to circle back around this, which was the best way to get work done before AI automation anyways.
13.
by piazz 5mo ago
Openclaw allows you to effectively “shell out” to another harness for your model calls, while still using Pi as your main agentic harness. This is the claude -p workflow. Tools and skills are injected into Claude and they hack session persi
14.
by piazz 5mo ago
If it's a big one and it's near you, you'd move away from the windows and heavy things that can fall, I suppose? For me I always just turn on iPhone screen recording and marvel at this amazing app and wish we had something li
15.
by piazz 5mo ago
Felt it all the way in Tokyo! There is this amazing app called NERV that, whenever there is a large earthquake anywhere in Japan, sends you an early warning push notification and an animated display with shockwaves emanating from the epicen
16.
by piazz 5mo ago
Yes, every day. Similar to other users here, giving it access to an Obsidian vault has been the key for me. And I wouldn't discount how much the chat interface matters - Telegram is so much nicer to use for extended conversations tha
17.
by piazz 7mo ago
Some creative workflows genuinely benefit from the tablet form factor. I often do serious photo editing on the iPad because I have access to Apple Pencil, and, somehow, holding the thing in my hands like an actual physical object activates
18.
by piazz 7mo ago
What an amazing feeling to see my flash animations I made when I was 13 on this site. Great project! What a unique era that time on the internet was. Can hardly imagine what my life would be today had it not been for Flash.
19.
by piazz 9mo ago
Thank you! Almost all of my customers so far have been directly from the central Anki plugin directory. I made sure to use lots of SEO friendly terms / buzzwords in the title so that when people ctrl+f for AI or ChatGPT, they find mine
20.
by piazz 9mo ago
Making almost exactly $500/mo on an Anki extension that embeds AI / text to speech / image gen deeply into the app, allowing you to generate example sentences, audio, explanations, etc, for whatever you’re studying, in bulk.
21.
by piazz 9mo ago
I do this, but it’s an Apple Notes file with a quick open shortcut mapped to the action button (side button) on my iPhone. I finally figured this setup this year. It had changed my life, in a minor yet significant way. (I also link to other
22.
by piazz 10mo ago
How is doomscrolling profitable for Apple? Frankly I think it’s the opposite - Apple is one of the only BigCo without an advertising based biz model. Unlike say Meta, Apple didn’t profit directly from increased engagement with your iPhone
23.
by piazz 10mo ago
You’re decreasing coupling at the cost of introducing more entities, and a different sort of complexity, into your system. Sometimes it’s absolutely worth it. Sometimes not.
24.
by piazz 11mo ago
I can’t explain why but “He’s assigning 128 to a string called Q” made me absolutely lose it.
25.
by piazz 11mo ago
This is good to know. I’ll probably play around with it sometime in the future. BTW, appreciate your many great write-ups - they’ve been invaluable for keeping up to date in this space.
26.
by piazz 11mo ago
Such a useful walkthrough. It looks like Mitchell is using an agentic framework called Amp (I’d never heard of it) - does anybody else here use it or tried it? Curious how it stacks up against Claude Code.
27.
by piazz 1y ago
Shameless plug for anybody who has been through the hell that is Anki card creation for language learning - I built an LLM powered extension for Anki that allows you to wire up fields to arbitrary prompts, and then generate notes in batch (
28.
by piazz 1y ago
> Another visitor was spending a leisurely week in Morioka alone before heading to Fukuoka for a few months. Oh hey, that was me! Small (internet) world, huh :)
29.
by piazz 2y ago
> Do a bunch of other things Curious about this! Currently trying to figure out a good 0.3mg melatonin dosing regimine for myself.
30.
by piazz 2y ago
Thank you :) And, sounds like great fun. Shoot me an email (in bio) or let me know how best to get in touch.

More ›