amadeuswoo
amadeuswoo on Hacker News
Excellent clear cut guide, will certainly share with my non-technical friends and family
Looks pretty good, the speech pipeline feels noticeably faster than the general-purpose apps I've used for lang conversations. Dead air kills immersion after all
Posted by
amadeuswoo on Thursday January 29, 2026 @01:14PM
from the show-hn dept.
https://suzerain.dev
I built a simple CLI that analyzes your Claude Code session logs locally and shows you how much autonomy you give the AI vs. how much you intervene.I've noticed on myself that bash command acceptance is where all the variance is. Every other tool (Read, Write, Edit) gets approved ~100% of the time. Bash is the one place people actually say no.Based on that signal, it classifies your style into...
Posted by
amadeuswoo on Tuesday January 27, 2026 @02:06PM
from the ask-hn dept.
Not using APIs, actually training a model from scratch, even a small oneWhat surprised you about the data, the training process, or the output?
The skills that atrophy are the ones you weren't using anyway. If you let the LLM do the interesting/engaging parts, that's on you, not the tool
The "management as superpower" framing assumes people thoughtfully evaluate AI output. In practice, most users either review everything (slow, defeats the speed benefit) or review almost nothing (fast, but you're trusting the AI entirely). The MBAs who did well probably had domain expertise to spot wrong answers quickly, that's the actual superpower, not generic "management skill
Very gorgeous and creative! Any plans to attempt other isometric cities? (Like SF, LA, Paris, London)
Yeah, the context window is a blunt instrument, everything competes for attention. I get better luck with shorter, more opinionated skills that front-load the key constraints vs. comprehensive docs that get diluted. Also explicitly invoking them (use the X skill) seems to help vs hoping they get picked up automatically
You don't, it's a map of intent, not infra state. What exists, why, what talks to what. Live state still needs IaC and observability. The .md captures the 'why' that terraform can't
One thing that’s evidently helped: using CLAUDE.md / agent instructions as de facto architecture docs. If the agent needs to understand system boundaries to work effectively, those docs actually get maintained
Honest question: has anyone found skills that fundamentally changed their workflow vs. ones that are just ‘nice to have’? Curious what the actual power-user stack looks like.
Anyways, great work on this btw, the agent-agnostic approach is the right call
Always appreciate an English major in the wild. But I think taxonomy is only wasteful if it doesn't map to real distinctions, good naming saves debugging time like when untangling "what did we mean by 'user' here?".
Wittgenstein said the limits of language are the limits of the world after all
It's worth noting Threads requires an Instagram account to sign up. That's like a 2B+ user funnel with constant in-app cross-promotion.
Not diminishing the growth, but "daily active users" hitting parity with X is a different achievement when you have that kind of distribution baked in Meta
Posted by
amadeuswoo on Monday January 19, 2026 @03:37PM
from the ask-hn dept.
Personally, I've been reading about how historical empires handled delegation and trust—who gets autonomy, who needs oversight, how that scales. Finding it weirdly applicable to how I think about system design and working with AI tools.Curious if others have pulled from history/humanities in ways that actually transferred.
Interesting architecture. Im curious about the workflow when an agent hits a denied action, does it get a structured rejection it can reason about and try an alternative, or does it just fail? Wondering how the feedback loop works between safety kernel and the LLM's planning
Stuxnet was 15 years ago. This isn’t crazy news, it’s just the first time it’s being reported openly
Nice! I’ve done some pitch detection with librosa before. Curious about your FFT/HPS implementation. How did you handle the time-frequency tradeoff for the spectrogram windowing? Or is that less of an issue with the other algorithms available as fallbacks
Don’t care if it’s vibe coded, love the raw and unapologetic aesthetic
I’m getting flashbacks on his 2009 deposition trial and something to do with “eggs”
Posted by
amadeuswoo on Saturday January 10, 2026 @10:02PM
from the ask-hn dept.
Recently built something where simple domain-specific heuristics crushed a fancy ML approach I assumed would win. This has me thinking about how often we reach for complex tools when simpler ones would work better. Occam's razor moments.Anyone have similar stories? Curious about cases where knowing your domain beat throwing compute at the problem.
The pipeline and incentive structure probably isn't there yet. I've hung out with game designers who are either bullish on AI but haven't integrated it, or just aren't tapped in at all.
Reminds me of the web3 wave a few years back, lots of people trying to make games with no actual understanding of what makes games fun and popular, not just profitable. The tech isn't the bottleneck, the taste is. Same applies to AI art
For mobile phones: Chinese brands (Xiaomi, OnePlus, Oppo) are worth considering. Comparable specs to flagships at half the price, and you're trading one form of data exposure for another rather than adding to it. Not "better" from a privacy standpoint, but meets the "not American" criterion and doesn't lock you into an ecosystem.
The feedback loop you describe—watching Claude's logs, then just asking it what functionality it wished it had—feels like an underexplored pattern. Did you find its suggestions converged toward a stable toolset, or did it keep wanting new capabilities as the trails got more sophisticated?
Posted by
amadeuswoo on Monday January 05, 2026 @01:56PM
from the show-hn dept.
https://nietzsche.amadeuswoo.com
I ran 5 English translations of Beyond Good and Evil through sentence embeddings to see if NLP could detect what I felt as a reader, that each translation reads like a different book.Findings:- Hollingdale sits at the semantic center, closest to the German (0.806) and to all other translators- Translators have fingerprints: UMAP separates them visually without being told who translated what-...