dasil003

Consider that there may be large populations doing both, but the overlap might be less than the internet jockey concludes via finger-in-the-wind hypocrisy check.

I don't think that nuance changes the big picture which is that Amazon has a conflict of interest between showing you the best product that you want, and showing you the product that the seller is most willing to pay to advertise. Quibbling over nuance just gives air cover to this fundamentally extractive rent seeking behavior. We really need stronger anti-trust action in the tech era given the extremely low friction and winner-takes-all dynamics of modern tech economy.

Correct. Don't get me started on Fidonet, as a young computer nerd in the 80s before even dial-up internet was available, that was a portal to the global internet that my adolescent shit-posting self definitely didn't deserve.

I vaguely recall those, but they were more for normies trying to get online. For me the old web is what I saw when I logged into my university gopher server and saw the advertisement for something called the World Wide Web which I could browse via lynx. Soon enough I got a PPP connection and then Mosaic/Netscape 1.0. However everything after javascript shipped (let alone CSS) is new new new. I'd almost go as far as saying if it doesn't have a tilde in the URL it's not old web... almost...

The problem is that the list of things one shouldn't do is infinite. The things one should do are narrow and get narrower along two independent axes of clarity/wisdom: system requirements and engineer seniority. LLMs overindex on any words given, so you only want to give them negative guidance around truly repeated, almost common-sense pitfalls. But in a large distributed system often the changes come from all different angles, and each agent will find it's own unique failure modes. Those comments will dilute critical context for diverse agents far more than they will help on average, at least for the systems I'm working on, given current claude code chattiness.

One way to clear some of the low-hanging slop is to just have a separate agent code reviewing and pruning any comments that don't stand on their own purely in the context of the diff, but it still doesn't catch it all.

I like the idea of using claude code in a separate git-managed repo. I do this for the management side of my job, and use Obsidian as the human interface on top.

However I would never use someone else's starter repo. The whole point is to augment my own perspectives and strengths. On top of that, the AI written pitch is extremely off-putting. If you can't even bother to take the time to write your own pitch, why would I believe for a second that your workflow has anything unique that goes beyond the lowest-common denominator AI capabilities.

When you are rich, it allows a lot more time and funding to come up with arguments supporting policies that will make you even more rich.

The best framing I've seen so far—and boy have a lot of people tried to frame the capabilities of AI recently—is that AI shines for people who use it to extend their thinking, and fails for those who use it to replace their thinking.

The latest frontier model developments have chipped away at that latter category. But ultimately the ceiling for that is mediocrity. Yes, if you imagine the most average common thing that has been done a million times, AI can absolutely give a bespoke version of it that matches your words of choice. For the moment it's still an impressive parlor trick to a lot of people, but the novelty wears off fast, and then you realize there's very little value in that. Doesn't mean everyone has to learn to code to use AI effectively, but it does mean they need to be curious and engaged or they'll just become part of the increasingly irrelevant slop machine.

I agree it's funny and won't really work on any kind of extended timeline. I mean Claude Code already added Auto-mode as a perfect example of this. But that said, I think it actually kind of makes sense in a transitional phase the power vs safety tradeoffs different users want to make varies so incredibly wildly that one product can't contain it all.

What I think will happen is that as model capabilities plateau (I'm not an accelerationist) the harnesses and products around them will start to specialize and they'll have different security models based on the product needs for those particular use cases.

For now, asking user to click a bunch of approvals, and occasionally making a mistake is a reasonable way to cover their asses until they see how bad security outcomes actually are in practice.

The basis of my comparison was nothing to do with business circumstance, but just that Apple and Oxide both were known for having taste in software (in Apple's case faded but still apparent in the DNA), and they build their own hardware to express that taste.

I think perhaps Solaris had a bit of that under Sun but I can't imagine that being the case under Oracle. And I'm not really an enterprise infra guy so can't really speak to IBM or Unisys offerings.

Sure. What's your point with regards to Oxide?

Being the "Apple" of enterprise compute (ie. vertical hardware software integration), solving the layers of aggregrate cruft that every single player in the Linux space (including Facebook/Google and other hyperscalers to some degree) have to deal with for compatibility since no one company has enough leverage to clean up poor or outdated legacy decisions in the stack. This is just a bigger market than anything EdTech, and with a real moat. Plus bonafides and actual evidence they are pulling it off, but that's honestly all gravy on top of the market opportunity.

It's funny you say this, because as someone that rejected emacs (default editor was XEmacs on the Sun machines in my 90s CS computer lab) in favor of vim early. I think a big part of the appeal of vim for me is that it's a great editor just based on the defaults. Sure you can configure it but you don't need to. Emacs strength is customization, and that's what Pi feels like compared to batteries-included harnesses.

I understand both world views and both are legitimate. But I do feel like LLMs are advancing so quickly that it's not a good use of my time to optimize harnesses at this point. I have actual work to do, so sharpening my tools needs to be selective and time-boxed. Personally I'm staying agnostic on harness, not locking into Codex or Claude Code, but also not prematurely optimizing things that tens of thousands of other tools-focused developers are going deep on across the ecosystem. My goal is not to be an early adopter but to reap the benefits of all that experimentation.

A lot of this rings true, but I think it's still too narrow. Sure, coding does not equal productivity, that is well debunked already. But I would argue that productivity is a product of engineering delivery + product decision making. Now where is the line between product and engineering? It varies by company, team and individual, but I don't think productivity can be measured for those functions independently, and in fact I see gains from AI on both the coding AND the product management side.

Basically as a senior tech lead in a large company engineering org, I don't have the bandwidth to individually validate every assertion from engineers on other teams OR from every product manager that comes with a half-baked ask. In the past I would be limited by the influence I could get through human relationships to strong SMEs with good judgment, and those folks always thin out as a company grows and calcifies. The number of creative and innovative thinkers dwindles, and the number of people protecting their turf and doing the minimum not to get fired increases. As a result many good ideas can get blocked by random gatekeeprs with poor imagination, poor expertise or both. However with AI I can follow up on gut instincts and fact check a lot more things, and ask incisive questions that can cut through a lot of organizational bullshit.

That's where I think most of the AI gains are today. Of course once AI plateaus and normalizes I think it will be baked into the org structures of tomorrow. But for now it offers real competitive advantage to those with the expertise to ask the right questions.

The only reason they deleted the static list of identified repositories was because of the negative PR of hitting HN. Although I have no affiliation with GitHub, I would imagine that no one with any resource allocation power is incentivized to take action to police public repos in general. The reality is that as soon as they start, it will become an arms race, and they will have to deal with false positives and the community headache that comes with that. They also have bigger fish to fry with the effect the influx of AI slop is having to their open source bread and butter. The two problems are not entirely disjoint, and I suspect solutions won't come until the problem gets much worse, and the agentic coding space stabilizes a bit more.

As a hiring manager I take candidate experience seriously because I know it’s easy to get wrong and leave people with a really bad taste in their mouth at a sensitive time.

As a candidate, though, I have learned not to take offense to being ghosted or the million other indignities that one is subjected to in a modern interview pipeline. It’s just a part of the modern economy, and far from the worst one. As terrible as it is, it’s not a hill worth dying on, suck it up and just move on with your search.

I get what you mean about the conflict of interest for consumer software, but I don’t think a product mindset is inherently toxic. It’s more a function of where you are in the enshittification cycle.

The majority of people who use LLMs today never even heard of ML though a non-trivial percentage have heard that modern AI is powered by LLM. You can’t forget what you never knew. Such is the evolution of language when a formerly niche technical concept crosses the chasm to mass awareness.

Is this a rant from an abused 19-year old? The lack of anything resembling nuance or recognition that the world is complicated renders any of the accurate observations sterile and unactionable.

For example, any who says something like this can not expect to be taken seriously:

> In my entire life, I have never known a parent that actually cared about his kids. Everyone had sent them to the torture chamber known as school, which to me, alone eliminates them

Some good points, but I think it still treats “tech” as too monolithic. It’s not simply that tech leaders made bad choices or that products are inevitably enshittified. If I had to summarize the root cause I’d say it’s that the scaling properties of tech have enabled capitalist and political excesses with a velocity and power never before seen, and it happened so quickly that we really haven’t had time to even understand what’s happening systematically, let alone update societal norms and governance to address the problems.

How can you separate those things? The changes happened because there were more people, and those people were valuable to market too. Hosting and moderation became more expensive, so that created a form of pressure as well. It's convenient to blame the producers (I hate Facebook as much as anyone), but I don't think it's terribly useful to try to hold them accountable.

I believe the economic forces were more or less irresistible. In other words, if the current powers that be had behaved more ethically according to early internet norms, the only thing that would be different is they would have lost in the market and been supplanted by equivalent mass consumer oriented companies pursuing the same enshittification cycle we dislike.

I don't think this can change unless there is a cultural shift away from worshipping at the altar of raw capitalism and GDP at the expense of everything else. The way our political discourse and regulatory capture have evolved recently I am not super optimistic, though I do think the mass hatred of AI across political lines does offer a glimmer of hope.

I'd wager that if a manager says that they want you to take it more like a real marathon and less like long hike.

Just realized zombo.com recently got a major update

You make it sound like a conspiracy but the article clearly argues that this is what the people want. If there really was that much centralized control then crypto would have been much bigger.

This smells more like the fluidity of what people mean by “junior” more than anything else. Journeymen engineers in their over-engineering phase, or even very “senior” expert programmers can suffer over fitting the product to their own mental model. The most senior judgment is to understand when an abstraction makes sense at a customer level, because that defines the durability of a business-logic abstraction.

Agreed, that's why I said used the phrasing "tend to". There's no silver bullet.

Huh? Hiring being broken has nothing to do with cost, it's a filtering problem. Even when there's no HR or bean counter in sight it's still hard. There's fundamentally limited signal you can extract from interviews, so there's very loose correlation to on-the-job performance. Saying it's a cost-cutting problem would just encourage more and longer interviews, which could actually work against you because high performers tend to have more options and will not jump through infinite hoops.

The call for CSS specialization strikes me as seriously misguided (even before agentic coding became a thing). Yes, web development is complex and you can’t know everything, but it is possible to learn what you need as you go and generally build up a strong high-level sense of what is possible and what approaches are reasonable for quality. But web development is not rocket science. Someone with CSS skills but not JS or probably some backend sensibilities as well is just not very useful. Being a software engineer inherently means comfort with change, it’s definitely a field that comes with a lot of challenge and potential for burnout if you don’t take care of yourself, but atomization of roles is definitely not the answer.

I think it was a joke

I don't really understand what you're nitpicking, the rule is "don't feign surprise". It seems perfectly well-stated to capture the spirit of the intent, and it explicitly allows for genuine surprise as you suggest.

Now, human interaction is squishy, so yeah, they are also trying to cover the all-too-common-in-tech case where someone is just being an asshole. Let's call it the Comic Book Guy case. In this case, it actually doesn't matter whether surprise is feigned or not, because what's actually happening is this person is listening and waiting for someone to express a blind spot so they can prove their intelligence by correcting them. You can't really write down an explicit deterministic rule for this, because it's all cognitive behavior social stuff that people are generally unaware of moment to moment. However the recurse center rules plus live feedback when it happens is as good of a solution as I can imagine.