nullc

42 ms·

451.
by nullc 1y ago
> maybe going forward we will be forced to come up with real solutions to the general problem of vetting people The monekeys paw closes a finger and now you need a formal certification, professional license, and liability insurance to pu
452.
by nullc 1y ago
> You can just reject the PR. And with a better and more useful response. Instead of wasting time on the technical details, you can give feedback like "this isn't the sort of change that AI is likely to be helpful with, though
453.
by nullc 1y ago
That's like saying "sorry, source code is too personal. In my 'open' project you get only binaries". ... and then I think about all the weights only "open" AI projects and walk off in disgust.
454.
by nullc 1y ago
FWIW, I can say from direct experience people that other people are watching and noting when people are submitting AI slop as their own work, and taking note to never hire these people. Beyond the general professional ethics, it makes you
455.
by nullc 1y ago
The person operating the LLM is not a meaningfully teachable human when they're not disclosing that they're using an LLM. IF they disclose what they've done, provided the prompts, etc. then other contributors can help them
456.
by nullc 1y ago
Better to think in terms of distrust rather than trust. Presumably if a contributor repeatedly made bad PRs that didn't do what they said, introduced bugs, scribbled pointlessly on the codebase, and when you tried to coach or clarify a
457.
by nullc 1y ago
I don't think it has anything to do with being a reactionary movement or counter culture. If it were I would expect, among other things, that it would absolutely prohibit the use of AI entirely rather than just require disclosure. The
458.
by nullc 1y ago
If their work was difficult to distinguish from AI then that sounds like a win too.
459.
by nullc 1y ago
> This whole original HN post is about ghostty btw Sure though I believe few commenters care much about ghostty specifically and are primarily discussing the policy abstractly! > because of the slot machine nature of it One could use
460.
by nullc 1y ago
We may be talking cross purposes. I read the grandparent poster discussing provably untested patches. I have no clue in ghostty but I've seen plenty of stuff that doesn't compile much less pass tests. And I assert there is nothi
461.
by nullc 1y ago
If the AI slop was that valuable a project regular, who actually knows and understands the project, would be just as capable of asking the AI to produce it.
462.
by nullc 1y ago
> is that they don’t know either It would be nice if they did, in fact, say they didn't know. But more often they just waste your time making their chatbot argue with you. And the chatbots are outrageous gaslighters. All big OSS p
463.
by nullc 1y ago
The utility of the rule is so that you can cheaply nuke non-conforming contributors from orbit when you detect their undisclosed AI use. Vs having to deal with the flood of low quality contributions on a individually reviewed basis.
464.
by nullc 1y ago
IJ FTW
465.
by nullc 1y ago
I used OpenSCAD a long time ago because it felt very natural having spent a lot of time with POVray as a teen. But I found it to be very difficult to design parts with specific dimensions that were needed to line up with objects in the real
466.
by nullc 1y ago
> slight mis-alignments during polar alignment For long enough observation there is no such thing as correct polar alignment on earth, thanks to atmospheric lensing. The density of the atmosphere shifts stars up with increasing amounts
467.
by nullc 1y ago
TBH I've found solidworks about as crashy as FreeCAD, okay maybe a little less but both are obviously crashy. FreeCAD has gotten enormously better (in terms of both crashy and general usability) in the last couple years.
468.
by nullc 1y ago
> 4. Walk away Run. Half phone will burst into flames.
469.
by nullc 1y ago
See my other comment: many do ask themselves that, in this case they reach the answer that it's a P&D and they think they can profit from it. > or whatever selling a course so some of those 'investor group' things t
470.
by nullc 1y ago
It's common for pump and dump to be extremely obvious. The reason for this is that many-- in some cases, virtually all-- of the victims/targets are aware it's a P&D and they believe they're going to get out early an
471.
by nullc 1y ago
My rates off-peak have increased faster than the on-peak rates, presumably because the on-peak rates were already sky high. In the last couple years peak has gone from ~0.36 to ~0.56, while off-peak went from ~0.12 to ~0.27.
472.
by nullc 1y ago
I think the multiple clipboards are great. My only regret is that there aren't more of them. Three would be excellent. But having multiple is very good, both for the pattern of "paste one thing over and over again, interspersed
473.
by nullc 1y ago
PG&E peak residential rates hit over 50 cts/kwh. PG&E works with a capture PUC and cost plus accounting where the only way for the company to increase profits is to drive up expenses.
474.
by nullc 1y ago
Why do you think the lack of time limits is an advantage? There is an amount of time spent gazing into your navel which is helpful. Less or more than that can be harmful. You can absolutely make yourself mentally ill just by spending too m
475.
by nullc 1y ago
You don't need a "regulated license" to hold someone accountable for harm they caused you. The reality is that professional licensing in the US often works to shield its communities from responsibility, though it's prima
476.
by nullc 1y ago
Stupid has boundaries on the evil it can do. Smart people of good intent are far more dangerous than stupid people.
477.
by nullc 1y ago
> or be monitored like you would be in a large city. Thanks to flock that's increasingly untrue. Most rural areas only have a few ways in and out. I've even seen roads closed off to force traffic past flock cameras. It's
478.
by nullc 1y ago
You don't mean to suggest that a scammy cryptocurrency entity that is currently bragging about attacking a competing system might ... lie to people???? Is that possible?
479.
by nullc 1y ago
> There is such a thing as useful proof of work. Not really-- or, rather, the security provided by proof of work is only proportional to the part of the cost above the fair value of the useful work. One of the main idea behind POW securi
480.
by nullc 1y ago
meanwhile the code in question imports os/exec and runs exec.Command() on arbitrary input. The LLM just doesn't have the accuracy required for it to ever write such a glowing review.

More ›