icedchai

17 ms·

541.
by icedchai 9mo ago
Non-routeable internal addresses are pretty effective at preventing external actors. When most people say "NAT", that is what they mean. You are technically correct in that 1) disallowing external actors is not a property of &quot
542.
by icedchai 9mo ago
Yes, but NAT combined with RFC1918 private addresses does provide a layer of security. This is the most common NAT configuration for 99.99% of residential users. It is what most people mean by "NAT." If your address cannot be ro
543.
by icedchai 9mo ago
Worse, there's sales in NOT doing it. When I buy a Mac, I get extra memory "just in case." I would've been fine with 24 gigs on my MacBook Pro, but I got 48.
544.
by icedchai 9mo ago
Back in 1993, I remember booting SLS Linux on a 386 laptop with 3 megs of RAM (1 meg on the motherboard, 2 meg expansion.) I could barely get it to startx and open an xterm, so I mostly used it in from the console! Before Linux, I was exper
545.
by icedchai 9mo ago
I get it, but these guard-rails are more suggestions and subject to interpretation. I would be more comfortable with a sandbox environment in a container. To be fair, I mess around with Claude Code and OpenCode running against various open
546.
by icedchai 9mo ago
It doesn't even need to go that low level, writing a program is just the worst case. There are ways to mass delete without `rm`. Example: `find` with the `-delete` flag.
547.
by icedchai 9mo ago
What if the LLM detects this, and chooses to run /bin/rm directly? Or worse, writes a program that calls unlink.
548.
by icedchai 9mo ago
When I see emojis in code, especially log statements, it is 100% giveaway AI was involved. Worse, it is an indicator the developer was lazy and didn't even try to clean up the most basic slop.
549.
by icedchai 9mo ago
I've had pretty good experience using Claude to "modernize" some old C code I wrote 30+ years ago. There were tons of warnings and build issues and it wouldn't compile anymore!
550.
by icedchai 9mo ago
I agree, but if LLMs are trained on common practices, best or worst, what do you expect? Testing, specifically, is heavily opinionated among professional developers.
551.
by icedchai 9mo ago
To be fair, many human developers are fond of pointless tests that mock everything to the extent that no real code is actually exercised. At least the tests are fast though.
552.
by icedchai 9mo ago
A coworker opened a PR full of AI slop. One of the first things I do is check if the tests pass. Of course, the didn't. I asked them to fix the tests, since there's no point in reviewing broken code. "Fix the tests." Th
553.
by icedchai 10mo ago
Thanks. It's fun chatting about this stuff! I don't hold mine strongly, either, though I am dealing with lots of AI generated slop code from others. Interesting times ahead.
554.
by icedchai 10mo ago
I don't believe that. If you go fully agentic and you don't understand the output, you become the manager. You're in no better position than the pointy-haired boss from Dilbert.
555.
by icedchai 10mo ago
You'd be surprised. I've worked at multiple startups where employees were given prod access with zero oversight on day one: AWS, sudo access, database passwords, everything. The one startup that didn't do that never launched.
556.
by icedchai 10mo ago
This is how we wound up with non-technical "engineering managers." Looks good to me.
557.
by icedchai 10mo ago
As an update, I gave in and installed 26.2. I've turned on "reduce transparency" and it's not that bad so far.
558.
by icedchai 10mo ago
Oh, I didn't realize this. I guess I'm screwed!! I'm stuck on 18.7.2...
559.
by icedchai 10mo ago
iOS 18 is still getting updates. Same with Sequoia. Eventually that will stop and I’ll have to upgrade.
560.
by icedchai 10mo ago
Very common. If you do get a number of out them, it is meaningless without understanding other details, like the liquidation preferences that may be held by current investors.
561.
by icedchai 10mo ago
I'm still on Sequoia on both my Macs. I updated my iPad Pro to iOS 26, decided it was meh, and am not updating my phone. My iPhone is over 4 years old and figure the new iOS will run like crap on it and then I'll be forced to get
562.
by icedchai 10mo ago
Without seeing a sample of the data, it's ambiguous. Example: It could be the value of an index fund.
563.
by icedchai 10mo ago
Whats your preferred local model?
564.
by icedchai 10mo ago
I'm not saying they can all do it now... but I don't think it's much of a stretch that they can learn it quickly and cheaply.
565.
by icedchai 10mo ago
This sounds nice, except for the fact that almost everyone else can do this, too. Or at least try to, resulting in a fast race to the bottom. Do you really want to be a middle manager to a bunch of text boxes, churning out slop, while they
566.
by icedchai 10mo ago
Yes, but the quality of output from open/local models isn't anywhere close to what you get from Claude or Gemini. You need serious hardware to get anything approaching decent processing speeds or even middling quality. It's m
567.
by icedchai 10mo ago
Many v4 prefixes in the ARIN region are legacy and don't support RPKI unless you sign the registration agreement. I have a legacy prefix and may eventually be forced to sign up.
568.
by icedchai 10mo ago
I worked on a Scala project about 15 years ago and it definitely felt overly complex. Similar to early C++, everyone used their own dialect.
569.
by icedchai 10mo ago
Most BGP peers have router filters in place. It's not 1996 anymore. I remember the days of logging into a Cisco connected to a Sprint T1 and seeing a coworker had fat fingered a spammer's route, sending it to null0. Oops. How did
570.
by icedchai 10mo ago
This matches my personal experience. I am primarily a backend developer. Can you recommend a book that teaches CSS the correct way?

More ›