Fishkins

8 ms·

1.
by Fishkins 1mo ago
> theoretically blocks any dangerous commands from running To be more precise, this article says auto mode blocked 89% of dangerous commands in their testing. A previous article[0] said it blocked 83%, but presumably it improved since th
2.
by Fishkins 4mo ago
> I missed the part where this has anything to do with saving time It lets you put the non-car stuff closer together, so you're traveling less distance to get to the same place. It requires urban design, not just a single person sw
3.
by Fishkins 4mo ago
The boat itself rocks, but do you see the background changing to indicate the boat is progressing through the environment? I only see that in the 3.1 Pro example. I believe that's what the OP meant.
4.
by Fishkins 4mo ago
Agreed. I think the approach described here is promising. Most of the workflow is deterministic and includes safeguards, but an LLM is invoked in the one case where it's really useful. https://lethain.com/agents-as-scaf
5.
by Fishkins 6mo ago
I'm sure most folks run Claude without isolation or sandboxing. It's a terrible idea, but even most professional software developers don't think much about security. There many decent options (cloud VMs, local VMs, Docker, th
6.
by Fishkins 6mo ago
Right, that's what I was referring to
7.
by Fishkins 6mo ago
> The part about permissions with settings.json [0] is laughable I never said "permissions", I said "sandboxing". You can configure that in settings.json. https://code.claude.com/docs/en/sand
8.
by Fishkins 6mo ago
Thanks for the data-based comment! Have you noticed any change in that trend in the past year or two, or is it continuing to get better?
9.
by Fishkins 6mo ago
I agree with most of this, with one important exception: you should have some form of sandboxing in place before running any local AI agent. The easiest way to do that is with .claude/settings.json[0]. This is important no matter how e
10.
by Fishkins 6mo ago
I do something similar. I leave up and down arrows alone, but have ctrl+p and ctrl+n behave as you describe.
11.
by Fishkins 6mo ago
> humans also make mistakes This is broadly true, but not comparable when you get into any detail. The mistakes current frontier models make are more frequent, more confident, less predictable, and much less consistent than mistakes from
12.
by Fishkins 7mo ago
Thus solving the problem once and for all https://www.youtube.com/watch?v=VW66EX75jIY
13.
by Fishkins 9mo ago
This is a couple of years old now, but at one point Janelle Shane found that the only reliable way to avoid being flagged as AI was to use AI with a certain style prompt https://www.aiweirdness.com/dont-use-ai-detectors-fo
14.
by Fishkins 9mo ago
I had the same experience as peer comments. I'm on Pixel 8 and Google Fi. When I check for updates, I'm told I'm up-to-date with the last update being over a month old.
15.
by Fishkins 1y ago
You should see an "unvote" or "undown" link to the right of the timestamp (i.e. the opposite side from where the vote arrows were). It's fairly subtle.
16.
by Fishkins 1y ago
Yeah, I never send a PR out without reviewing each commit myself and adding GitHub comments when I think it's relevant. Sometimes a PR is clear enough that I don't feel the need to add comments, though.
17.
by Fishkins 1y ago
I'd say "good old days" thinking is probably involved, but not the full explanation. Over the past few decades, software has gone from a fairly obscure profession to being seen as a great way (maybe the best way) to make a lo
18.
by Fishkins 1y ago
For others who didn't know, the -u flag in the OP's command makes it so ripgrep _will_ search files even if they're gitignored
19.
by Fishkins 1y ago
It isn't nearly as big of an issue as the phone size, but it is still a nuisance. I know there's no chance of it ever coming back, but I'd like it to. I still have a small amount of hope that someone will make a modern, well
20.
by Fishkins 1y ago
Yeah, the main news I want to hear is the release of smaller Pixel phone. Secondarily, I'd like the return of the 3.5mm port. I don't care about any of the stuff they actually announce. I do currently use a Pixel, but I hate how b
21.
by Fishkins 1y ago
I'd say a large part of the country had the same sentiment about George W Bush. I'm not sure whether that was true or just an act, though. In politics, I think the opposite of Hanlon's razor has often been applicable. It&#x
22.
by Fishkins 1y ago
At one point the org I work for used it for internal documentation. They migrated off of it many years ago, though.
23.
by Fishkins 1y ago
Huh, that's interesting. Mixing indexes and FKs is a major conceptual error. FWIW, I've also asked everyone I've interviewed in the past decade about indexes and FKs. Most folks I've talked to seem to understand FKs. The
24.
by Fishkins 1y ago
Thanks for making that PR! A regular `brew install git` installs 2.50.1 for me now.
25.
by Fishkins 1y ago
Early on at my first full-time job I requested a week off to attend a Quaker conference. My manager approved the request without comment. After I returned from vacation he asked how the competition went. He'd never heard of Quakers and
26.
by Fishkins 1y ago
You're on the right track with money, but that's not exactly how I'd describe the goals. My impression is the primary goals are getting things coded and billed properly and thoroughly, and getting things documented in a way t
27.
by Fishkins 2y ago
> Ah, but - in large and poorly-maintained codebase, it is always the case that some refactoring is already in progress, likely stalled for weeks, months or years as priorities have shifted without it having been completed. This is true,
28.
Yubico Issues Security Advisory as 2FA Bypass Vulnerability Confirmed(forbes.com)
10 points by Fishkins 2y ago|0 comments
29.
by Fishkins 2y ago
Custom feeds and starter packs seem to be promoted much more on Bluesky than lists were on Twitter, but how different are the actual capabilities from Twitter lists? I never looked at my algorithmic feed. Instead I make a "list" o
30.
by Fishkins 2y ago
According to the Mozilla docs, O(n) lookups would actually violate the spec [0]. As a sibling comment says, a tree set is one option to satisfy the spec. Another is a linked hash set, which would have O(1) lookups [1]. 0: https:/&#x2F

More ›