actsasbuffoon

I was the first dev at my current company to experiment with Claude Code back when it first came out. Some of my coworkers tried it, and some didn’t like it at all.

But now literally all of us are using it. The company gives us a $100 monthly stipend for it. We’re a small dev team, but our CEO is thrilled. He keeps bragging about how customers are gobsmacked by how quickly we’re delivering new features and he’s been asked if we’ve made a ton of hires in the last year. We’re actually down two developers from when I started.

I don’t love the code it writes, but the speed is unbeatable. We still need devs, and I don’t think that’s ever going to change. But we don’t need as many devs. We’re shipping like crazy with a small team. I don’t think more people would speed us up much at all.

> I can't think of a single time where having someone else review my work or give me feedback is a meaningfully bad thing

I have ample examples, unfortunately. I had a coworker whom I liked as a person, but had a nasty habit of using PRs as a way to hijack all decision making. He’d leave PR feedback for his personal preferences and mark them as “changes requested.”

Everyone feels like an asshole giving an approval when someone has requested changes, so you had to comply with all of his preferences if you wanted to merge your code.

I’ve had several companies where I submitted a PR and had someone say, “You’re guarding against an edge case that won’t happen. This is over engineered. Remove it.”

And it made it less than a week in production before we hit the edge case that I’d been forced to neglect.

I had a team come to me with a request, so I built it. They were thrilled. Then another engineer was like, “I don’t like (some technical detail). You need to change (major architectural decision).”

I gave the re-architected version to the team who requested it and they said, “Wait, I loved what you built before. What is this? I don’t want this!”

This post resonated with me pretty hard. Hire good people and deputize them to make decisions. You’ll end up with something good more often than not. I’ve never seen design-by-committee produce a great product.

I’ve had too many experiences with seeing decent contributions get worse and worse as they go through successive rounds of feedback.

As a musician, this happens to me as well. If I hear a piece of music in a noisy environment my brain will fill in the gaps. I’ll think, “Wow, this music is really interesting.”

Then I’ll hear it in a quiet environment and realize that I preferred the version where my brain adding things to make up for what I couldn’t hear before.

I guess it’s a little like diffusion. The brain has natural denoising processes which, in an unconscious way, taps into our tastes.

Before the ACA, insurance companies were allowed to have these things called “lifetime limits.”

Basically, once your healthcare got expensive, they could just cut you off and say they wouldn’t cover you any further. And because of pre-existing conditions (which the ACA also eliminated), you couldn’t get new health insurance. You were basically fucked.

My mom got cancer a few years before the ACA passed. So far as I’m concerned, the old insurance system killed my mom when she was only 40 years old. I lost my only surviving parent, and my little brother lost his mom when he was only 10 years old. So forgive the utterly flabbergasted look on my face as I read your comment.

So many distortion pedals use an op amp to run a signal into antiparallel diodes to create distortion. I’ve spent a few weeks trying to emulate it, and it’s a lot of fun.

Different flavors of diode make significant changes to the way it sounds. Even things like LEDs can be used (they are Light Emitting Diodes, after all).

Andy Simper of Cytomic is some kind of mad genius at this stuff. He’s created a painstakingly accurate emulation of the Ibanez Tube Screamer that allows you to change the values of basically every component in the circuit diagram. It’s jaw dropping: https://cytomic.com/product/scream/

He’s also shared a ton of incredible information about how he emulates circuits. The math can get really intense. If anyone is looking for a fun project, I strongly suggest experimenting with circuit modeling. It’s a great workout for the brain.

AKA the culture of “Why is AWS down again?”

This is a delightfully horrible idea. Well played.

I’d go so far as to say that you and this other engineer both fill vital roles on your team. It’s not that you’re right or he’s right; you’re both right.

I’ve been on teams where I needed to be the methodical engineer who carefully built critical infrastructure and agonized over every decision. I’m currently at a small startup that hasn’t yet reached breakeven, so I’m scrambling like crazy to build things our customers and investors will pay for. That’s what this team needs.

Thank goodness for both of you. Your team would be worse for it if they lacked either.

Considering how much harm he did to farmers in his first term, and they overwhelmingly voted for him again in 2020 and 2024, I’d say they apparently prefer this over the alternative. I don’t expect this to change anything.

I’ve been waiting to see my countrymen vote in their own best interests for my entire life. I’d be shocked if they started doing it now.

We are astonishingly bad at understanding abstraction. We understand that if Kevin shows up at our house and punches us in the face, we should avoid Kevin. But once you put a layer of abstraction between Kevin and the broken nose, we suddenly become baffled about how this could happen, and then we vote for Kevin.

I agree with the overall sentiment. It’s all just a bunch of functions with well defined inputs and outputs that can be combined in a delightfully endless number of configurations. None of it is nearly as complicated as it seems at first glance, and learning how to do it is a ton of fun.

No, the other commenter is correct.

There’s a gate signal, typically activated by holding a key (though in a modular synth, there are many other potential gate sources). While the gate is open an ADSR envelope progresses from Attack -> Decay -> Sustain. It then remains on Sustain until the gate closes, at which point it enters the Release phase. So the amount of time it remains in Sustain is dictated by the gate signal. Notice there’s no G in ADSR, because the gate doesn’t come from the envelope.

What you’re describing is Hold, which some envelopes (AHDSR is one popular flavor) can do. Many Elektron groove boxes have hold stages on their envelopes.

In AHDSR, an open gate goes from Attack into Hold, where it retains its value for a set period of time after the attack, and then goes into the Decay phase and continues from there.

There are plenty of other kinds of envelope, and things that live somewhere between LFOs and envelopes called a function generator, which is often an AR envelope that can be looped. Then there are complex many-stage envelopes that were especially popular when digital synths were first coming onto the scene.

I’d also add that the description you gave of a synth architecture is generally considered East Coast synthesis. One or more oscillators going into a mixer/VCA, and then into a filter and possibly some effects is a very popular architecture made famous by Bob Moog. But there are other forms like West Coast synthesis where instead of having filters, you run gentler wave forms like a triangle through a wave folder, and/or a complex oscillator where you have a pair of oscillators that can cross modulate one another. So where East Coast synthesis takes harmonically rich waveforms and cuts harmonics away with filters, West Coast synthesis starts with harmonically tame waveforms and adds harmonics through various flavors of FM, wave shaping, etc.

Then you’ve got samplers, granular synthesis, physical modeling, additive synthesis, and a bunch of other types as well. The East Coast architecture is popular, but there’s a lot more out there.

It would be ridiculous if it were true. The methodology here seems a bit muddled and I’m not confident that they’ve measured what they think they’ve measured.

Many people do perceive a difference in how upsetting a death is. For example, we’ve got two people. Kevin killed an able-bodied adult male. Tom killed a 3 year old girl.

Based on nothing more than that, do you dislike Kevin or Tom more? Maybe I’m not supposed to have a preference here, but I’m definitely feeling angrier at Tom.

The same is true with men and women. There was a rather famous case with John Ellis, the executioner. He’d executed many men and he’d been okay with it. But when he was ordered to execute Edith Thomas, she was so terrified that she collapsed and had to be dragged to the gallows. He resigned the following year and had multiple suicide attempts, and was ultimately successful in taking his own life. He killed people for a living, but killing a woman affected him more deeply. This was over a century ago, so I don’t think this had anything to do with wokeness.

Meanwhile, studies have shown that people tend to view things as being more dangerous when they’re tied to immorality. For example, when asked how dangerous it is for a doctor to leave her child in a car for 15 minutes while she performs lifesaving CPR on an accident victim, people thought that was safer than the woman who left her child in a car for 15 minutes to have sex with her affair partner. Obviously the danger to the child is the same, but our moral judgement is different. So questions about ICE agents are probably colored by that. If a poor African villager doesn’t feed their child because the have no money or food, is that better or worse than an ICE agent who doesn’t feed their child because they’re too busy arresting people for “looking like an illegal immigrant?”

Is there some kind of bias being measured here? Yes, I think so. But I’m not confident that it’s even remotely comparable to what the author claims it is.

Maybe. This assumes I trust Microsoft to have part of my computer where I have no ability to interrogate it to see what they’re doing in there.

If it’s on my computer, I should be allowed to read and write to it. End of story. I don’t care if that makes it vulnerable. So far as I’m concerned, letting Microsoft keep secrets from me on my own computer is similarly catastrophic to losing my HD to a crypto-locker virus.

Correct, this bill won’t pass. But the weirdos keep introducing these kinds of bills, and they’ve already won in a bunch of states. I consider it to merely be a matter of time before it happens at the federal level.

And once it becomes law, it’s very difficult to get those rights back. Politicians don’t want to be on the record fighting for anonymous access to obscene materials.

Yeah, I can’t explain why this project makes me so happy because I struggle to think of any time where I’d need this, but it puts a big, dumb grin on my face.

I disagree, at least for the next 5 years or so.

Even with humans, it’s very common to ask them for something and get something back that’s completely different than what you wanted. Short of reading your mind, AI is going to require a lot of info to get the desired result.

So someone is going to have to gather requirements and break them down into a clear, well thought out set of instructions for the computer. That’s literally what programmers already do. We’re just programming in a different language now.

Completely agreed. I’ve got about 20 years of professional software development experience. Tools like Claude Code let me build incredibly quickly when I use a stack I know intimately.

But when I try it with a stack I don’t know, I quickly find myself needing to learn the new stack to get anything done.

Extremely experienced engineers will be able to move quickly with these tools, but specialization will still exist, and I’m not clear on how juniors are going to ramp up on all of this.

It’s funny you should mention LISP macros. Five years ago, I’d have told you that the most efficient way to build something is to get an extremely experienced dev with mastery over a flexible language like LISP, and just let them go nuts. I constantly have to hold myself back from metaprogramming tricks because my coworkers won’t be able to maintain it. But on my personal projects, I really get to flex my muscles.

It occurred to me a few months ago that that’s exactly the same spot in my toolbox where Claude Code is. It’s a tool that allows one person with mastery of a tech stack to build things incredibly quickly. Look at Cursor where a tiny team has over $100M in annualized revenue.

Big companies often find themselves spreading ownership across too many hands. You know exactly how to build what you need, but you’re not allowed to touch 80% of the code involved because it belongs to a different team. But coordinating all of that is a massive amount of overhead that will derail most projects. This is why so little happens at big companies.

But these tools enable a smaller number of experts to do a job that used to take many. I expect the speed improvements to be super-linear as a result.

Of course, that brings concerns about what happens to the employment situation for devs if that happens. I hope it means that 10,000 new startups can build more ambitious products and absorb all the people who will probably be laid off by the huge corps in the next few years.

I regularly have this happen:

1. Ask Claude to fix something

2. It fails to fix the issue

3. I tell it that the fix didn’t work

4. It reverts its failed fix and tells me everything is working now.

This is like finding a decapitated body, trying to bring it back to life by smooshing the severed head against the neck, realizing that didn’t bring them back to life, dropping the head back on the ground, and saying, “There; I’ve saved them now.”

The original definition of vibe coding meant that you just let the agent write everything, and if it works then you commit it. Your code review and security check turned this from vibe coding into something else.

This is complicated by the fact that some people use “vibe coding” to mean any kind of LLM-assisted coding.

“Agentic coding” is probably more accurate, though many people (fairly) find the term “Agentic” to be buzz-wordy and obnoxious.

I’m willing to vibe code a spike project. That is to say, I want to see how well some new tool or library works, so I’ll tell the LLM to build a proof of concept, and then I’ll study that and see how I feel about it. Then I throw it away and build the real version with more care and attention.

This mirrors a weird thought I’ve had recently. It’s not a thing I necessarily agree with, but just an idea.

I hear people say things like, “AI isn’t coming for my job because LLMs suck at [language or tech stack]!”

And I wonder, does that just mean that other stacks have an advantage? If a senior engineer with Claude Code can solve the problem in Python/TypeScript in significantly less time than you can solve it in [tech stack] then are you really safe? Maybe you still stack up well against your coworkers, but how well does your company stack up against the competition?

And then the even more distressing thought accompanies it: I don’t like the code that LLMs produce because it looks nothing like the code I write by hand. But how relevant is my handwritten code becoming in a world where I can move 5x faster with coding agents? Is this… shitty style of LLM generated code actually easier for code agents to understand?

Like I said, I don’t endorse either of these ideas. They’re just questions that make me uncomfortable because I can’t definitively answer them right now.

I literally just had a conversation with my CEO this morning where he told me not to disclose the projects I’ve been working on, so I can only speak about it obliquely.

We identified some problems our customers have, and I’ve come up with interesting ways to use LLMs as part of an automated system to solve some of those problems. It’s not the kind of thing where we just dump some data into the ChatGPT API and get an answer. We’re doing fairly deep integrations that do some interesting/powerful things. It’s been a big deal for our prospective clients and investors.

That’s the thing. There’s value in AI, it’s just not worth half a trillion dollars to train a new model that’s 0.4% better on benchmarks. Meta is never going to get a worthwhile return on spending $100M on individual engineers.

But that doesn’t mean AI is without its uses. We’re just in that painful phase where the hype needs to die down and we treat LLMs as what they really are; an interesting new tool in the toolkit that provides some new ways to solve problems. It’s almost certainly not going to turn into AGI any time soon. It’s not worth trillions. It’s certainly worth something, though.

I think the financials on developing new frontier models are terrible. But I’ve already built multiple AI projects for my company that are making money and we’ve got extremely happy customers.

Investors thought one company was going to win the AI Wars and make a quadrillion dollars. Instead it’s probably going to be 10,000 startups that will build interesting products based on AI, and training new models won’t actually be a good financial move.

At a bare minimum there needs to be some way to understand how close you are to these limits. People shouldn’t be wondering if this is going to impact them or not.

Pornography has never been easier to access and teen birth rates have been falling for decades.

Once puberty hormones kick in, teenagers are going to start figuring out that there’s something going on. We’re far better off educating them about the risks involved rather than letting them blindly experiment.

Pornography is probably the worst way to accomplish this goal, of course. Comprehensive sexual health education is vastly preferred. But pretending sex doesn’t exist is how you end up with pregnant kids.

The bulk of the harm from ultra-processed foods was specifically from meats, with smaller contributions coming from sugary drinks and dairy desserts. It’s the pink slime, reconstituted McRib, and hot dogs that are causing the most significant health problems.

Beyond burgers have no cholesterol, hormones, or antibiotics. They’ve got significantly lower saturated fats. Studies have shown that swapping out regular burgers for Beyond burgers lowers your LDL cholesterol and TMAO.

I’m not going to pretend they’re as healthy as a burger made out of black beans and carrots. But if concerns about UPFs are your primary reason for avoiding them then you can relax; they’re not that bad.

There’s a very simple fix for this. Don’t make H1B a lottery system. Grant them in order of highest-paying roles.

We get access to truly exceptional people for whom companies are willing to pay exceptional wages, and we eliminate the exploitation of H1B visa workers.

Imagine hiring a person to do this job at your company. They show up and behave the way the LLM agent behaved in the article.

Not only would the person be fired quite quickly, but people would be telling stories about the tungsten cubes, the employee inventing stories about meetings that never happened, giving employee discounts at an employees-only store, and constantly calling security. It would be the stuff of legends.

I worked at a company where there had been one outrageously overworked employee who had finally been pushed too far. He shoved his computer monitor to the floor and broke it. He quit and never returned. They were still telling stories about that incident almost a decade later. I’m not even sure the guy broke his monitor on purpose; I wasn’t there, and for all I know he accidentally knocked the monitor over and quit.

So if that’s the bar for “insane behavior” for a human, Claude would be the kind of legendarily bad coworker that would create stories that last a century.

I try to just give honest feedback, and I don’t feel badly about it or worry about blowback.

For example, I applied at a company for a fairly high ranking position and did really well. The technical co-founder said I gave the best performance he’d ever seen on their code exercise. The internal recruiter said I was the first candidate he’d ever given the maximum score to.

Then I had one round where the interviewer showed up very late and was instantly rude. Would barely even talk to me. He had this disdainful look on his face for the whole interview.

The internal recruiter called me after and was basically like, “What the hell happened in that interview? Everyone loved you and said you were amazing. Now Chuck (not his real name) said you’d be a bad hire. What is going on?”

They still made me an offer, but the whole thing left a bad taste in my mouth, so I went elsewhere. But apparently a couple of pretty high ranking people were pissed that I walked away, and there were some consequences.

Should I feel bad about that? Honestly, I’m glad. Their interview process was broken, and I hope I improved it so this shit doesn’t happen to anyone else.