jknoepfler

A healthy measure of my personal skepticism comes from the fact that GenAI represents the convergence of so many historically negative (and frankly idiotic) influences.

- AI hype-mongers. We've dealt with these for decades in academia and industry, through countless boom/bust cycles. Yes, there's progress, but reality is boring and selective.

- Tech Bros. NFTs? The MeTaVeRsE? The CEOs of OpenAI, Anthropic, Meta, and SpaceX have proven repeatedly to be dishonest grifters chasing all and only money. They're like the tobacco industry of laughably dishonest at this point (and consistently wrong).

- Singularity grifters. The Kurzweil cultists flooding the ranks of lesswrong feel suddenly emboldened. They've been issuing batshit proclamations and navel gazing about AI alignment to anyone who will pay attention to them (and write them a check) for decades. It's no more connected to reality now than it was then.

- The LinkedIn hivemind. If you aren't transforming your organization to meet the new GenAI future, your company is BEHIND and you are FAILING AS A LEADER. Be the one firing white collar workers, not the one being fired. Make sure you look good on paper with bullshit performative AI initiatives or your ass is dust. Oh, no ROI? No problem. Read my book on NAVIGATING THE ENTERPRISE AI TRANSFORMATION. It's skin-deep and filled with empty platitudes and clip art, but it will make you look like a big brave tech leader, just like Steve Jobs or Dario.

- Investors. Yes, we're in a bubble. Yes, we've been in a bubble before. Yes, it's going to pop. Yes, that's going to be bad. Really bad, probably. Remember the financial crisis? Investors don't. And this time, there won't be a deep-pocketed government to bail anyone out, because debt is already sky-high and we're still recovering from COVID.

So while I get there's something there, and I'm keeping abreast of it and extracting the best and leaving the worst, it is an exhaustingly negative experience. My guard is up and I have absolutely run out of patience for bullshit. It's not funny or smart or interesting. It's not even fun to complain about. The forces in the mix are awful and I don't want anything to do with them.

If a craftsman injures themselves or a co-worker with a faulty tool, the tool manufacturer is very often liable for damages. I struggle to see GenAI any differently.

Please take 5 seconds to fact check yourself with a very quick query against your information source of choice.

Over ~95 degrees, a fan blowing on a human acts as a convection oven, increasing body temperature while dehydrating the body.

This is greatly exacerbated in extreme heat.

CDC guidance is to avoid using a fan in extreme heat conditions unless it is combined with a cooling source and careful hydration. Once again, not exactly an option in a Texas prison.

Fans have no cooling effect once the ambient temperature exceeds human body temperature, so "making do with fans" is literally making do with nothing.

Those citizens "making do" generally have freedom of movement, free access to cool water, air-conditioned businesses or vehicles, and are not locked down in a concrete facility.

There is a absolutely no nuance to this.

That's a very dodgy etymology. The use of "slop" to refer to trashy content dates back to the late 19th/early 20th century, long before the advent of the internet.

I think our tendency to attribute "novel" word usage to 4chan might reflect more on us than on our language.

I'm a little short on primary reference material, but here's an example from 1917: https://www.gutenberg.org/cache/epub/22692/pg22692-images.ht.... Conversation on language log has at least two others from 1897 and 1920 respectively.

72% of job postings on company career sites are closed within 90 days. Pretty sweet.

Hard disagree about "more intelligent". They are playing a financial game using a combination of leverage and reptilian ruthlessness to EBITDA hack.

They buy one (or more) companies, often with only the slimmest understanding of what those companies do, slash opex by gutting the company with layoffs (yay EBITDA), maybe staple a few such companies together with leveraged buy-outs, then resell the whole bundle for more than they paid.

From experience, they don't give a single crap whether the resulting mess still functions. They care about selling the company for more than they bought it for. That prospect is only tenuously and at best accidentally related to whether the company still functions.

The private equity companies I've had to deal with were full of braindead MBA spreadsheet monkeys and used car salesmen. Their chief differentiator was that they worked 80 hours a week and were enthusiastic about laying people off without much deliberation.

Thinking private equity is "more intelligent" than business owners is like thinking house flippers are "more intelligent" than home owners. No. They know how to rip out carpet and replace it with laminate on the cheap. They know how to cut corners and hide it. They know buyers will over value a fresh veneer of paint. They don't give a shit about the long-term health or value of the house. They are not better stewards of houses. They specialize in short-term profit maximization and that is literally it.

They have suffered from abuse because of unregulated police. They have not suffered because of police simpliciter.

Police perform an important function. Unregulated, unchecked, untrained, militarized police taught that they are immune from prosecution for crimes against citizens do not.

Drawing an equivalence between these two muddies the waters in an extremely unhelpful way.

A well-regulated police force drawn from the population it is sworn to protect is a good thing. It's a hard, important job.

And yes, it's obvious Flock was going to be abused and no I don't think a well-regulated police force should have access to pervasive surveillance. But that doesn't mean I want to do away with police in general.

The whole ACAB narrative is one of the most deeply unpopular views of the American left, polling at something like 80% disapproval. I guess ACAB people think my friend who went to the police to report sexual abuse by her step father (leading to his arrest and incarceration) should have hired a... community councilor? A mercenary? Gotten out the community pitchforks? Because that ends well as a mechanism for justice? I genuinely don't know, but I think people who want less policing (as opposed to accountability, transparency, and demilitarization) aren't thinking very clearly.

Why would it require explanation? It was stupid beyond belief to begin with. It would be like rescinding a policy profiling soldiers using phrenology or waving some scientology wand over them or something.

The fact that people think there's anything even worthy of discussion on this is alarming to me.

Or maybe people like to stare at the clouds while herding sheep. This is just bro-science half-baked evolutionary psychology bullshit.

Personally I enjoy listening to the birds with a book in my hands.

The solution is very clearly neither deregulation nor defunding regulatory agencies.

I sign in to video calls using my browser every day. Saying video/audio requires an app beggars belief.

Privacy is the opposite of what you get with an app. Privacy? Really?

Local storage is available to browsers. Office 365 literally functions in a browser with full access to local storage.

Anything processor intensive can be compiled to WASM. This isn't 2005 anymore.

So... widgets?

I can see an argument for games, where you want to pre-install a large amount of resources, and maybe have a better i/o story. Even that is a function of convenience over necessity.

Did domain names stop functioning all of the sudden? Going to somedipshitsbar.com/menu is about as complicated as entering a wifi password at a coffee shop.

I guess this will be a good test of the open source model DuckDB has opted into.

As I understand it, DuckDB is MIT-licensed and governance is provided by a separate, non-profit DuckDB foundation. This was done explicitly to prevent BSL-ification.

I'm very curious to see whether that approach holds the line. I wouldn't bet on it, but one can hope?

Nothing has made me want a downvote option for HN posts more than the endless AI advertisement slop that has flooded the site.

It's a familiar story. If you copy/paste Wikipedia and turn it in, you learn nothing. If you skim Wikipedia, hit up the references, the consult other sources, then synthesize your own thought, you probably get farther faster than you would without it.

We could probably cook up thousands of examples of the same problem (YouTube DIY tutorials, GPS navigation, etc.)

This is known as the "Paradox of Automation." It's a real problem for many industries.

The OG paper on this is called "The Ironies of Automation" (Bainbridge, 1983). It's a clunky but fascinating read. I give it a skim every couple years and usually come away with a bit of a fresh take on the problem.

Workers without 5-10 hours of unpaid commute time in their week are happier. Gee whiz.

That's time to run errands, learn a skill, exercise, sleep, read, do nothing... it's almost as though our time is valuable and we don't like wasting it.

Which I worried was so obvious it wasn't worth remarking... but here we are.

I'm sure you know this but RSUs are such a fraction of a percent of real world compensation that they might as well round to zero.

That's the plain truth, and I think it's unfortunate that we need to use sympathetic examples as a bridge. I guess I've accepted the premise that the privileged gestalt is profoundly distorted and that "charismatic macrofauna" are a necessary ladder to be kicked away.

The example is true, but to your point, isn't representative... and that doesn't matter.

Honestly this isn't the right conversation to be having at all.

In the United States, ~40% of the homeless are employed. A deeply non-trivial percent either are kids or have kids. Everyone should volunteer at a support center for homeless youth some time.

Sure, there may be some kind of Pareto principal in play for the hardest 20% of cases to crack. Who cares? Let's eliminate 80% of the existing problem and worry about the rest when we get there.

I'm going to go out on a limb and suggest that the median HN reader has very little appreciation of the human reality of homelessness in the United States, and primarily see it through the lens of people at the extreme, visible margins. I know I used to until I got involved.

edit: I mean all of this in a very boring, matter of fact way, not some kind of holier-than-thou virtue signaling way. The 80% problem of homelessness is a tractable economic problem. While it's true that the United States has an almost dizzying level of propaganda demonizing poor people, it's also true that real homelessness is hard to spot because people have a natural instinct to preserve their and their family's dignity. The first time I came to terms with this was coaching debate in highschool. I had a varsity debater who couldn't afford pants to meet the dress code. Incredibly bright, academically successful kid. His father was an unemployed teacher working menial jobs, his mother was taking care of three kids. They lived in a tent at campsites. I had absolutely no idea. We obviously bought him pants and paid all his travel/etc. fees (he'd been making excuses to not go to events requiring travel). That's a very boring, normal case.

That family should have had a free or nearly free housing option. They did not. They still wouldn't.

Yeah, if you offload work to someone else and provide direction, it's called management. You know who I wouldn't trust to maintain and operate production software systems? Me, the person managing multiple teams. I wouldn't know what I was doing and I'd make a mess of things.

You know who I definitely don't trust to maintain and operate production software systems? the coding agents my teams use.

As a manager I require my employees to understand what they are building and maintaining as engineers, not as managers. There is no viable alternative to that today.

Somehow the American AI industry managed to create a product I trust less than existing commercial offerings. Kudos, that is actually pretty impressive. Considering the stakes, one might also call it foolish. But I'm sure they've done their market research, they've identified their product niche and they're racing to meet it at the bottom.

<rant>

y'know, I am sincerely trying, in good faith, to let my teams guide me in how best to intentionally leverage GenAI. I'm trying to use it in good faith myself. I really do reserve the right to be wrong about it.

It's just... the problem was never the ability of GenAI to produce code that works on a superficial level. The problem was always that GenAI would push engineering teams at companies yet further away from what they were building. Generations of compilers, interpreters, frameworks and ORMs have kept pushing engineering teams out there. Cloud infrastructure has pushed engineering teams out there. Giant data analytics platforms have pushed engineering teams out there.

Engineering teams need people who love learning about how things work, but those people often have a rough time in the industry. People who are passionate about software and systems engineering, who love learning how systems work, who love diving deep into niches of reality... those people have been getting hit in the head since they entered the work force.

They're essential, and they've had decades of this crap to learn from. History has taught them (us) that businesses are really, really good at opting in to bloated, obtuse, unmaintainable solutions if some sales monkey can convince leadership that they no longer need to hire smart people who speak a language they don't understand. And they (we) are used to dealing with the consequences. Every single experiences engineer has their war stories about it.

Speaking personally:

I had a partner who worked at a startup in Seattle. They were a novice programmer. I saw what happened when their employer tried to build a Ruby on Rails stack with only junior engineers out of college. It ended horribly.

I worked as a first hire on a clean-sweep purge of a .NET team that had forgotten how to program. The company was bloated and nobody was accountable for anything for years. Then one day they get a mandate to actually build a thing for a big new online initiative. Nobody knew what they're doing. Wrappers around wrappers around wrappers of copy-pasted stackoverflow code. The company hired a hard-ass director to reboot things because nothing was getting done. He mandated a total shift to Golang/Kubernetes over .Net/some managed Microsoft stack, and said "you'll prove you can do it or you're out". I was hired because I could. I was substantially younger than everybody, but asked to write architecture docs and establish design patterns and sit on hiring panels for all the new devs. Because I don't know, I was a nerd who learned how things work and gave a shit and had somehow picked this knowledge up. They fired 8/11 of the devs on that team, as well as the project managers and the dev manager. We grew a new team behind it that was 1/3 the size and moved immeasurably faster. I hate everything about how that whole story played out. I'm not proud of that chapter of my life. I resent in particular the leadership who allowed engineering fundamentals to erode over time because it wasn't their problem.

I've worked at companies that fled to the cloud from local infrastructure. They did it with this almost insane desire to stop paying people to solve complex infrastructure problems. Balls-out, fuck these nerds who actually know how hardware works, they're weird and we're sick of them, we're going to pay Amazon a billion dollars to never have to hire them again. And then they hired some (often quite competent!) devops folks to migrate everytrhing... those people did... things got wildly expensive... meanwhile businesses locked in to stupid technologies like DynamoDB and Lambda because some insufferable sales idiot connected with a CTO in a sweat-yurt... and et voila, here we are writing the proposals to unfuck our cloud migration and return to local infrastructure. Yay CapEx. Boo OpEx. Venture Capitol batteries possibly included. Just gross incompetence by a generation of LinkedIn leadership buffoons who don't know the first fucking thing about how computers work.

I've been there. You've been there. LLMs are one more way of being there. We feel it in our bones. We have noses and we can smell. Trust was lost long ago. This shit is old. And we're really tired of it.

And this time it's especially annoying because it's big and tied in with an economic bubble and companies laying people off and general major disgruntlement. It's a really bad vibe right now for a technology leader, and if you're running a company I hope you get it, at least a little bit.

I really give a shit about being a good leader in technology and I'm really, really frustrated right now because there an almost un-navigable divide between what I believe to be true (from my experience) and the narrative I'm expected to sell. That's a hard position to be in. But I (and my peers) try our best to occupy it with grace.

</rant>

I hire contractors for a large enterprise in the US. The going rate is typically $85-$100/hr for a senior dev, depending on specialization. Lead-level maybe $120 for the right skill set.

No thank you. I had ICE flood the streets where I live with poorly trained thugs with assault rifles recruited from the ranks of neonazis illegally assaulting, detaining and abusing citizens of my country because they had brown skin. They shot dead two citizens (Renee Good and Alex Pretti) doing what I was also doing - relentlessly recording and documenting their behavior and addition to protesting.

I don't think you realize how ridiculous and out of touch it sounds to tell me to "disconnect" from politics to be less angry. I am extremely angry, and I will continue to let that anger fuel my choices until the vile bullshit that has infected this nation is gone.

Enlightening my ass. Tell that to the trodden upon and the dead. Reality sucks and sometimes you have to do something about it or else you're the problem. Tough shit if you don't like it.

<rant>

I manage several teams of developers who use them every day professionally and use them for personal projects privately. Just last week, at the prompting of said devs, we had a working agreement conversation about curtailing the use of AI in our codebases because of rapid erosion of our teams' ability to operate, update and maintain codebases that had started to spill over with slop.

We have had multiple incidents of credential leakage, integration tests wiping live databases, comically broken code that passed vibe-written tests, documentation and code comments that were hallucinated and/or fake, and most importantly developers saying "we no longer know how this code works but it's massively bloated and unreadable and we can't tell you with a straight face that we can maintain it or fix it if it breaks." We have seen a flood of vibe-PRs from engineering adjacent teams that suddenly think they can code shipping prototypes into production that do not work do not scale and cannot be maintained. I am personally writing the tickets to decom one of those today. Which is great, I love telling business "the progress that was reported to you was a lie, this shit never worked, don't shoot the messenger but also don't let this happen again."

I embrace GenAI as a productivity tool for people who know what they are doing. It's a +10-15% velocity boost. That's great! That's a big deal, devs are expensive, and that might push some kinds of business model over the threshold into viability. That's great!

It is not, however, transforming the profession as I know it, it is rather making my job harder and less pleasant and it is making my leadership dumber by the second.

Absolutely no comment on vibe business decisions / vibe OKRs / slop reports or the host of other garbage that has started to creep into professional life. I have had to have some very uncomfortably direct conversations with peers in leadership about using complete bullshit to make decisions, and it is very, very frustrating. Do you know how hard it is to convince someone that metrics their bot hallucinated don't exist and would be meaningless if they did? You can't convince someone of something they are incentivized to not understand. It's been very eye-opening in terms of who I can trust to actually make sense when it matters. I'm grateful for the clarity.

Meanwhile we are rapidly losing brainshare from the top because our principal/staff engineers are pissed off and have the bankroll to just leave. We aren't hiring and training younger engineers to keep the talent pipeline moving. Which boy howdy is THAT going to cost us unbelievable sums of dollars to fix in the medium-term future.

I am in the uncomfortable role of trying to make the best of this but it would be a metric ton easier if the narrative from the c-suite aligned with reality in any meaningful way.

</rant>

No, their internal PRs have the same problem, but they are pushing an AI narrative from the c-suite and can hide the reality of their private code bases.

The "limited ROI on AI investment" articles will continue to percolate slowly into the brains of the LinkedIn hive-mind until we hit a tipping point, and then we'll finally shut up about how a handy dev tool with some decent use-cases is the dawning of the singularity that will replace all white collar labor and get back to actually building business value.

Let's just pause for a second and steep in the irony of "we're all in on AI, except where there are visible consequences."

The same problems that bedevil AI-written PRs here bedevil AI-written PRs at my place of work.

I'm sure it'll get better eventually, maybe, but something something Pareto principle.