You can choose to not wire LLMs up to a gun.
15,316 karma · joined October 17, 2013
Enough. Let's take them at their word: it's time to sue them, jail them, stop them from killing us all like they insist they are trying to do.
"Defendants admit that they provide a service without fully knowing how it works. This is not just any service. It is one that Defendants themselves concede poses an existential risk to the continued survival of humankind. Defendants claim they cannot stop barreling forward with their potentially civilization-ending endeavors unless they are forced to do so by the government. They have asked the government to tie them to the mast. Plaintiff brings good news to the Defendants: The Florida Attorney General is answering your cry for help with a motion to enjoin you from harming Floridians with your reckless, unacceptably risky product."[1]
[0] https://www.theguardian.com/technology/2026/oct/03/openai-sa...
[1] https://www.myfloridalegal.com/sites/default/files/plaintiff...
Or, alternatively, the comment you're responding to is NOT empty — there is value in reminding all the AImaxxers on this board that GPT v10 may not have one-shotted the end of human suffering by early 2027.
For the developers here, consider: "So our CI pipeline is super flaky, and it you have to restart it in a variety of non-intuitive ways. LLMs have been a massive breakthrough because the bot can run in an endless loop reading the confusing error messages and retrying the pipeline with different environment variables until it passes!"
To use the metaphor from the article, they found the keys in the last place they looked.
> The equities here could not be more one-sided. Defendants admit that they provide a service without fully knowing how it works. This is not just any service. It is one that Defendants themselves concede poses an existential risk to the continued survival of humankind. Defendants claim they cannot stop barreling forward with their potentially civilization-ending endeavors unless they are forced to do so by the government. They have asked the government to tie them to the mast. Plaintiff brings good news to the Defendants: The Florida Attorney General is answering your cry for help with a motion to enjoin you from harming Floridians with your reckless, unacceptably risky product.
If AI levels out most other distinctions, maybe in the end we choose to give money to people we trust.
Now they're trying a new thing. I don't know if it's so much "being bold" as flailing around for some new thing. They really desperately want to own a "platform" like Microsoft, Apple, or Google.
This is a hilarious premise if you work in a domain where it matters even a little bit whether the data is correct or not.
I could just as easily argue that by paying my ISP and signing up to their T&Cs, I've opted in to seeing their ads, not the ads from some random website.
You mean like how Google makes money showing ads against your content that you don't control? You mean how basically every ad network works?
I think software development is a field unusually vulnerable to LLMs because there's lots of good examples for training data, and the work is easy to validate in short order according to somewhat-objective criteria.
Note that certain kinds of mathematical proofs also fit these criteria, and also have been an early showcase for AI abilities. Anything where the bot can get to a good result by just grinding on it for days is going to get taken over by AI quickly.
You and I have very different definitions of "engineer". To me, an engineer is a licensed professional who adheres to the code of practice for his or her profession. We'll have to agree to disagree on the semantics. I think lots of software people want to call themselves "engineers" because it sounds like someone who should be paid more money.
...or it did, until the AI revolution.
Has any software "engineer" ever lost his license because he signed off on trash code that failed? Where is the Otis elevator emergency brake[0] of the software world?
[0] https://www.eng.auburn.edu/~kchang/comp6710/readings/They%20...
[1] https://www.otis.com/en/us/tools-resources/high-rise-safety-...
I'm very very curious where all these mythical software engineers are working. Finance, maybe? ... or not https://www.bloomberg.com/opinion/articles/2025-03-03/citi-k...
>> One quirk of the program was that the amount field came pre-populated with 15 zeros, which the person inputting a transaction needed to delete, something that did not happen.
> One quirk? Fifteen zeros? The quirk in the software is that if you type “send $1” it says “okay sending one quadrillion dollars” and you have to backspace over all the zeros? Incredible interface design.
Do you actually write software for money?
Most commercial software is never designed, it evolves, over the course of repeated "agile" "sprints". Every two weeks, someone hacks together something that kinda-sorta works and then that is released.
Eventually (often many years later) the thing becomes so complex and unwieldy that it's impossible to add any more features, so someone gets angry, throws it away, and starts again growing a new monster.
> They simply pass messages back and forth.
So you claim. We're about to conduct a real-life experiment to see which is more valuable: passing messages back and forth to the computer, or passing messages back and forth to the customer.
[0] https://en.wikipedia.org/wiki/2024_CrowdStrike-related_IT_ou...
I think there will be a convergence where AI enables more POs to create software directly, and more SEs move up the value chain to work on product-level problem solving.
What did you think "disruption" meant? Vibes, papers, new Javascript frameworks?
Strongly disagree with this. In contrast to most jobs, software engineering is unusually self-contained, with short feedback loops and immediate validation. You (or Claude) implements the feature, you run the tests, you fix the bugs.
Most white collar jobs are nothing like this: lawyers have to prepare briefs (AI can help here) but they have to file briefs, they have to wait weeks or months for rulings, they have to argue in court. Marketing execs might spend millions preparing a new campaign, launch it for the holiday season, and it has to be right the first time. Most white collar work involves judgement, anticipation, and ambiguity in a way that software doesn't.
Software lets you guess-and-check over and over again until it works.
Good pay, low barrier to entry, pleasant. Pick two.
Healthcare is another growing field that's easy to get into with only 1- or 2-year certification programs for some positions.