scruple

I use coding agents and LLMs at work where I'm more or less to some degree required to. At home, I write code the old fashioned way. Not katas, etc., necessarily, but I've decided that if we live in a world where code is cheap (or cheaply generated) that I'll hone my Lisp skills. I haven't used a Lisp in a few years and it's brought a lot of the joy of programming back for me, at least at home.

What percentage of people "seem accepting of violence as a tool" when they're starving to death?

I guess? Yes. But I played for like 20 seconds and lost interest.

> I still have a lot of usage for AI: Exploration, Double-checking me, teaching me.

I'm ready to give up on having it even review my code at this point. It's been so frustrating. It hallucinates bugs, especially in places where "best practices" are at odds with reality.

Recently it informed me of a bug where it suggested the line of code in question couldn't possibly do anything because on Linux the specific stdlib behaved in X ways, but it was obvious from the line of code that it was running on Windows which doesn't have this problem at all. Of course, it doesn't actually mention that this is an issue on Linux, just that there is a bug here. It vomits up a paragraph of $WORDS explaining why this was a high-priority bug that absolutely needed to be fixed because it was failing in subtle ways. Yet the line of code in question has been running in production, producing exactly the results it is expected to, for ~3 years.

And this is just one simple example, of the many dozens+ of times it has failed this task this year. In that same review run, the agent suggested 3 additional "bugs" or other issues that should be addressed that were all flatly wrong or subjective. I'm at a point of absolute exhaustion with this sort of shit. It's worse than a junior half of the time because of how strongly opinionated it is. And the solution to this sort of problem is an endless amount of configuration and customization that will be forgotten about by all of us over time, leading to who knows what sort of knock-on effects (especially as we migrate from one model to the next). We have a guy on our team who has ~17,000 words in his agent and instructions files, yet he sees nothing wrong with this. I guess he just really loves YAML and Markdown.

Yeah I write prompts asking it to misspell a few words, break a few grammar rules, forget to capitalize once in a while, miss some punctuation once in a while. No one will ever catch on.

LLMs are Brandolini's Law taken to an entirely different plane of existence.

I started my career in embedded at an RF company. Back then, I was 20 years junior to the next oldest guy and he was 20 years junior to the rest of the engineers. It was an incredible place to start, learning from some crusty old veterans who were pushing into retirement age. I ultimately left because the pay wasn't there. I've often thought I'd love to go back, even if it meant a decrease in pay, because the environment was so rich with learning and experimenting.

Their point is that the jobs problem or flourishing is a mechanical necessity for maintaining the Mandate and surviving whatever things come next.

> Is this something that people are seriously thinking they'll be able to do and successfully stay gainfully employed and contributing to the world?

No. I resisted for a bit but have started using it at work. Mostly because I believe usage is now being monitored. I'm in a very high-scale engineering environment involving both greenfield and massive brownfield codebases and the experience is largely a net loss in productivity. For me and some others who I've spoken to in my org, opting in is a theater that we're required to engage in to keep employment and not a genuine evolution of our craft.

These tools struggle with context once you get deep into a codebase with many, many millions of lines of code and sprawling dependencies. Even for isolated Python scripts or smaller, supporting .NET apps, the time spent correcting subtle bugs or bullshit, or just verifying the bullshit, often exceeds the time it would take to have written it from scratch.

Regardless, what I've observed is that these tools do nothing for the actual bottlenecks of software engineering: requirements gathering (am I writing the right thing?) and verification (does it work without side effects?). Because LLMs are great at generating text, they're actively exacerbating these issues by flooding our process with plausible looking noise.

> I know I should be used to people openly lying with no consequence, but it still amazes me a bit.

Well that makes two of us. Character seems to mean nothing today.

> Can someone help me to understand why OpenAI and Anthropic talks as if the future of humanity controlled by them?

He wants to build the AI that makes people's lives better. Okay. Did the people ask? Do they have a say? It's all very easy for a billionaire to say when it's just him and a couple of people in his cohort in the driver's seat.

Beyond that I'd like to simply know why he thinks any of this is his responsibility. It seems much more obvious to me that he simply found himself in the right place at the right time and is trying to seize it all for himself as if it's his to take.

Incendiary. Is he trying to suggest the journalists are at fault here?

Yes. This administration is on a blitz ahead of the midterms. Rubio has his eyes on Cuba and he's positioned himself next to Trump. It is happening.

It would be absolutely astounding to read TFA and then continue to take this weasel and his word at face value.

Aren't these things just legal fictions that only exist because/as long as the US government allows it?

So it's a controlled demolition? In this view, is the goal to devalue our own debt? Extract value while it still has value?

Service for citizenship, stationed along the border walls and manning really big fucking guns seems to be the place my brain always goes to in these sorts of conversations.

Through circumstance I went from a fully remote working arrangement to working for a place that's 5 miles down a single surface street from where I live (which, during commute hours pushes up to 15 minutes). I go in twice a week, but since we have kids it's mostly a wash. I still vastly prefer the 3 days per week that I work from home. The office gets very busy and very loud, especially in the afternoons.

My employer very successfully shipped a lot of brand new stuff during COVID when everyone was full-time remote. We made a lot of money off of those products. Then they sort of bragged about it. Then they instituted RTO. Now (I'm told, I'm at a megacorp so I'm far removed from these discussions) the executive team is bitching that we're not in office 5 days per week.

What's the basis for this war in Iran? Did that stop this administration? This is akin to pointing out that it's actually illegal to drive 30 mph over the speed limit.

I'm far more interested in understanding how we can 10x our confidence in a change and not just our line counts.

How much more productive are we supposed to be in engineering? Are we 10x'ing our testing capability at the same time? QA is already a massive bottleneck at my $DAYJOB. I'm not sure what benefits the company at-large derives from having the typing machine type faster.

In the end it's a distinction without a difference. The result is the same, and Thiel was building the fire escape before anyone could smell smoke.

> Since you don’t care that much about money, they can’t buy you at least.

I get the sense that Bill does care about money, and so does Larry Summers, so Mom's advice probably wouldn't have done much there.

Having 1 place (or just generally limiting them) that does the things keeps the dry_run check from polluting the entire codebase. I maintain a lot of CLI tooling that's run by headless VMs in automation pipelines and we do this with basically every single tool.

It is exhausting and frustrating. What level of evidence is required here? And why? It's not unlike the article. Do we need to wait around for firing squads or something? What does it take for people to wake up?

Is this common belief grounded in reality? We're well past the notion that the "old man" (and his VP, as well as most, if not all of his cabinet) believes in the law and he has stacked the deck in his favor.

ACC generally has a 3-4 second time interval that it permits between you and the car in front of you. I live in SoCal, so a lot of my driving is on very aggressive routes. The 4-second gap is mechanically safe but it's practically unusable because it creates a void large enough to invite other cars to lane change in front of me. So when that car merges in, the ACC detects a violation of the safe braking distance and decelerates to reestablish the gap. I call it the "cut me off" loop when we're on trips.

And before anyone suggests that I start tinkering around with the settings, I have adjusted it and the damned thing just resets itself constantly.

Probably more like 90 minutes, tops.

Probably not. I love what I do, I get to work in games and it's so much fun. But at the end of the day I still have a issue, in trading my time for money, and there's a lot of stuff I'd rather be doing. Spending more time outside, exercising, reading, cooking, being with my family. Work is... Just work.