ebcode

5 ms·

1.
by ebcode 8d ago
> In present reality, AIs don't do anything a human doesn't tell them to do. This seems unlikely to change. Not in my reality. I ask AIs to do things (write code, mainly), and they do. But sometimes they go ahead and do somet
2.
Graduating into Disruption: Labor Market Outcomes for AI-Exposed Majors [pdf](www2.census.gov)
1 points by ebcode 10d ago|0 comments
3.
by ebcode 11d ago
Mostly I was thinking about environmental damage, honestly. There's plenty in the historical record about damages from nuclear testing. https://en.wikipedia.org/wiki/Starfish_Prime https://storymaps.arc
4.
by ebcode 12d ago
While I have my reservations about Amodei and his company, I'm nevertheless a happy user of their software. And I'm in agreement with him (and Sanders) that we should all. slow. down. To my mind, the last great arms race between n
5.
by ebcode 16d ago
Thanks for asking the question. Preface: I do think the nature of the work is changing, now that LLMs are capable code generators. And I'll admit that I started working with LLMs because of FOMO, but also to keep my skills sharp in an
6.
by ebcode 16d ago
Let me propose: HNmind.
7.
Ironies of Automation (1983) [pdf](ckrybus.com)
3 points by ebcode 18d ago|0 comments
8.
by ebcode 19d ago
Culture flows from the top. Bill doesn’t take responsibility in his email, so he’s teaching everyone downstream to pass the buck like he does.
9.
by ebcode 22d ago
From https://code.claude.com/docs/en/remote-control : Feature-flag evaluation: DISABLE_TELEMETRY, DO_NOT_TRACK, CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC, and DISABLE_GROWTHBOOK each disable the feature-f
10.
by ebcode 26d ago
I'm only running a few agents in parallel so far, but I don't let them write or read anything w/out my approval. This slows things down quite a bit, but it means that I have exactly the opposite of your problem of stopping th
11.
by ebcode 1mo ago
SQLite doesn't have system-versioned / temporal tables, but a quick search turns up a fairly straightforward approach. Instead of overwriting existing rows, always write new rows with timestamps. https://www.ohnekontur.
12.
by ebcode 1mo ago
Our government is tyrant.
13.
by ebcode 1mo ago
It’s interesting to think of the LLM as a kind of “mirror”. It reads to me like this person is engaging with the tool in a fairly hostile manner, and the LLM is like, “so you want to play it like that, huh? Well, let’s play it like that the
14.
by ebcode 1mo ago
My personal opinion: all output from LLMs is hallucinations. The idea that it’s “right” or “wrong”, when stating factual information is really in the eye of the beholder/user. What year did astronauts land on the moon? — is a question
15.
by ebcode 2mo ago
I don’t trust their outputs yet enough to not review everything myself. But I’m happy to be. a QA bottleneck in that way. What I want now is orchestration to get them running in parallel. I’ve looked into Yegge’s beads, but I want something
16.
by ebcode 2mo ago
I mean, it does feel like we could pause here for a few years, just to give everyone time to adapt to what the current models are capable of. We've seen back-sliding with model releases (anecdotally, but maybe there's published re
17.
by ebcode 2mo ago
Between those two, it's Claude Code. Gemini models are behind in the benchmarks.
18.
by ebcode 2mo ago
I’d say the problem is that, at least in the US, the government was conceived as a democratic republic, but private enterprise wasn’t required to follow the same conceits, and so all our corporations behaved more like small dictatorships. N
19.
by ebcode 2mo ago
Can we get that added to the US Constitution and the UN Declaration of Human Rights?
20.
by ebcode 2mo ago
What if a single email cost $0.001 cent to send, and it was paid to the recipient? For $10, you could send 10,000 emails. For recipients, every 1,000 emails they get is a dollar in their wallet. You’d need something like a blockchain for th
21.
by ebcode 2mo ago
Have you tried VirtualBox? That’s my go-to for running un-trusted code. Start with a clean VM/fresh install, and don’t put any secrets in it.
22.
by ebcode 2mo ago
check out David Ackley's work
23.
by ebcode 2mo ago
I tried it on my iPhone but can’t scroll the spreadsheet. still it seems like a fun project.. will check it out on desktop when I get home
24.
by ebcode 2mo ago
Here's the original blog post from last December: https://ebcode.com/blog/2025/12/17/sourceminder/ The 15% number is based on two runs from the ./experiment folder, on an instance from the
25.
Show HN: Sourceminder.org - token-efficient code search(sourceminder.org)
4 points by ebcode 2mo ago|2 comments
26.
by ebcode 2mo ago
It’s a tie.
27.
by ebcode 3mo ago
https://wettel.github.io/codecity.html
28.
by ebcode 3mo ago
And it reads to me like they have some other reason to move on from SWE Bench Pro, but they don't want to say what it is. They say right up top, "~30% of the tasks are broken." But that leaves ~70% un-broken, which seems pret
29.
by ebcode 3mo ago
Still plugging away on SourceMinder ( https://github.com/ebcode/SourceMinder ). I know, I know, everyone and their brother is working on token-saving schemes for LLMs. But I’ve found it to be useful even without the LLM.
30.
The Great Software Quality Collapse (2025)(techtrenches.dev)
2 points by ebcode 5mo ago|0 comments

More ›