mekoka

I can’t conceive of the leadership of the Python or Rust or any other community I’ve ever worked with doing something like that.

I'd be prudent to use a clear faux-pas by a BDFL as an argument to push for an alternative, seemingly more consensual, leadership style. I'd also be careful to fall for the apparent lack of overt aggression in the latter type of structures, as a necessarily positive signal. I've seen people in various such communities (including the two cited languages) that have chosen not to interact with their steering committees, because of perceived toxicity.

This can and does happen with any kind of structure and can take many shapes, some of whom are too subtle or passive to even accurately pinpoint, in the way you would for example directly index a BDFL. In The Tyranny of Structurelessness Jo Freeman made the case that the lack of a clear structure in the make up of a movement presents with opportunities for very fuzzy power dynamics and related abuses that are also much more challenging to deal with.

As entertaining as that article was when it came out, I think its apparent wisdom should have been reevaluated for at least the past two decades. Even back when it was written, its pertinence was questionable, since it hinged on the perception that written code has particular value, to justify much of the failures at Netscape and Borland, neglecting many of the other -- probably more relevant -- business and human factors that were at play within the two organizations and their surrounding ecosystems. But lessons learned from watching startups fail in trove in the following two decades have mooted many of Joel's arguments. With the vantage of hindsight, if you read JWZ's account of what went on at Netscape, some of the dysfunctions become glaring.

Making software projects successful has always been about a lot more than just writing code. People have been "rewriting code from scratch" successfully even before LLMs. We just don't tend to call it that. We call it "cloning", "competition", "copy", "alternative-to", "reverse-engineering", "x-written-in-language-y", etc.

Short publication intervals. That's the heuristic you're going with? I have dozens of such unpublished long form articles on fundamental topics, just waiting for me to carve some time to polish. I expect that on a sabbatical I could finish and post them in quick succession. It's an unreliable indicator to determine AI-writing.

Being dismissive is one way to think yourself "better" than others. But on HN hastier reactions tend to be outright negative or out of context with long articles, as eyeballs actually evaluating the content will take some time before providing constructive criticism.

How were you able to pick this up? Not challenging your assertion. Just really curious. Can you point to some clues? I read it (with my own eyes). I can't see actual evidence the text itself is artificial or at least, that it is not human-curated.

Ignorance will always be a better starting point for discovery than wrong assumptions. If you leave comments, they must reflect what the code is actually doing. If during edit it's no longer the case, at least mark them as stale. The next best thing is indeed to remove them.

The article had great advice on this actually. When asked how to get into C, or whether it was more advisable to learn Zig instead, Mitchell suggested that

It’s more important to learn how computers work and make the language just a means to understanding how they work. [...] even in this age of higher level abstractions and web development, it’s still important to understand the basics of CPU scheduling, memory, cache hierarchies, file systems, disc and file access. When you work directly above the syscall layer, whether in C, Zig or Rust, it really helps you understand what’s happening[...]

If this is the ideal model, why don't authors skip the middleman themselves and just put the .pdf/.epub on their website directly in exchange for donations?

On the same note, why do game devs need to give Steam 30% of their money and not just sell to the public directly and pocket the 30%?

We're needlessly making this into a general problem. Why hastily discuss ideal models? The current model is fine and the issue isn't generalized. We're talking about having the option to skip asshole middlemen, or to be more specific, Amazon. A company so big that solving this special case on its own leaps us a huge portion of the way into solving the problem at large.

Is the general sentiment that Steam is also an asshole?

Using an ORM should not exempt you from designing your schema around the domain. If anyone thinks it's the case they're either using the ORM wrong, or using the wrong ORM, or perhaps creating the wrong ORM.

Similarly, designing your schema to match the domain does not necessarily grant you the productivity boons of an ORM.

Having (ab)used Postgres with and without ORM, I've never had a situation where the latter imposed any kind of design decisions on the schema. They're orthogonal concerns. Itself, the choice of using an ORM tends to be motivated by experience with certain requirements in the business logic. I love SQL, but when having to deal with API resources and their various representations, marshaling, validation, options, etc, it's difficult -- and to say the least, impractical -- to stay principled to the "no-ORM" and "raw-SQL" mottos.

You stopped right before the best part. When they decide to create their own ill designed, badly tested, and undocumented mapper.

you give up so much in performance.

Not really. ORMs (memory) and databases (disk) are distant by multiple orders of magnitude performance wise. Skipping the ORM to shave off some cycles is akin to haggling over a few pennies on your thousand dollars bill.

It’s doing more harm than good.

For context

Can we PUHLEASE stop this AI Confidence Theater, people? It’s doing more harm than good.

The word "entertain" quickly came to mind in that sentence because attracting the audience as a primary concern is typical of the entertainment industry, regardless of the content being educational, news, or otherwise. If we get stuck on that choice of a word, it's admittedly a bit narrow for the point I was making, but I think the rest of my post successfully got things across.

what they want to watch

What does that even mean? Not everything published has to be with the purpose of entertaining. This is a more recent mindset that came with the advent of the "Content Creator" (think Mr.Beast). But it needs not be this engineered. Historically, some of the more interesting channels on YouTube have been people organically sharing something that was genuinely dear to them. Offered with little concern as to whether it's "what people want to watch". They make you want to watch it, because it's a passion that they manage to convey. By contrast, people who are particularly concerned about eyeballs seem to publish with the purpose of meeting the audience's interest as a main guiding factor. And yes, maybe for those, YouTube is a better fit. But I also know many for whom moving to a reliable alternative (tech wise) would be a step up.

If for instance your goal is to share educational content for free, provided that the hosting platform is sound, your main concern as far as your audience finding your content should be that of discoverability. Which is an easy enough kink to iron with current search technology.

I expect the documentary to get an uptick in downloads as well.

This isn't a specifically American phenomenon. Humans have fear buttons. If you press the right ones, a sizable portion of any population would trade in significant amount of freedom to feel safer. Those for whom it's almost pathological spend considerable effort trying to pass on their worries to others, or chasing opportunities to exert some satisfactory control. If they succeed in either goals, you get less freedom.

It's not really about going back. Evolution happens within a pattern of ebb and flow, back and forth. We never get anything perfectly right. We overdo, then course correct, rinse and repeat. Right now, we're embracing AI, but we're also noticing atrophy of skill as an effect. These may be the last generations of such craftspeople that can notice, compare, and inform as to whether there's actual loss. That future you're seeing for yourself is still being written. Stay tuned.

Just try to take those same friends who drive everywhere on a (walking) trail hike inside the city, then come back to report.

It depends how you store and look things up. There are so many optimization opportunities and strategies to make the lookups fast that this is pretty much a non-issue in practice (e.g. deny list implemented as any combination of runtime or in-memory index, trie, or bloom filter). At the first invalid bit the lookup fails and the token is allowed to proceed to subsequent auth checks. Which should happen quickly for the vast majority. No need to head straight for the worst possible implementations, like whitelist disk lookups.

Every once in a while an article with a sensational title against JWTs pops up in here and I have to wonder if something new was discovered. But nope. It always boils down to the same "can't invalidate it" complaint, which can be addressed with a viable deny list structure (simple in-memory or runtime index, bloom filter, trie). Then there are the vague "it's insecure" claims which when looked at closely, come down to "some people use it wrong".

I haven't read the book, but I've thought about similar problems. Sharing my attempt at a solution.

First, the organization must have an original steering body made up of "true believers" in a shared mission and vision. This body must make preventing the short or long term erosion or dilution of its mission a priority. Meaning that there's awareness that this particular corruption is a possibility from the start. Also, an understanding of the typical mechanisms through which it can happen (usually very human).

Second, as part of its survival strategy the steering body has the responsibility to identify other genuine true believers and integrate them in its makeup. There must be built-in assumptions that many outsiders will be attracted to its powers of influence and will try to infiltrate it.

Joining the steering council should thus be done primarily based on "culture fit". Admittedly a rather segregating practice, but one which in this case comes with the advantage that certain signals are just hard to fake on the long run. So, although many could try to dress, look, talk, or walk the part for a while, there will always be some shibboleth that trips up impostors.

I foresaw some problems to this structure that I haven't yet worked out, but it feels like a step in the right direction, if I had to come up with a solution.

It would be true if they bothered hiding it. But as the featured author said, people seem increasingly not shy of simply forwarding you a screenshot of the AI answer.

Your analogy is apt in more ways than one. It comes down to how often the point of a journey is to get to the destination. Most old wisdoms teach that the latter is more often just a MacGuffin to embark on the former. If they're right, AI offers tremendous potential for new adventures, but also as a catalyst for completely missing the plot. Yes, we're "free" to choose, but I'm skeptical that a culture conditioning us to eschew friction necessarily equips us to distinguish when the grind and frustration might be "good" for us.

I once made a travel friend who just didn't get the point of me taking eight hours to slow travel by train or by bus across a country that we were both visiting, when she could just hop on a plane and get to the next city in an hour. Earlier in my youth, transportation choices were economically motivated, but what I got from it would influence all future visits to other countries. When chilling with other travelers, exchanging tips and stories, it was as if my friend was visiting a completely different place. She left the country shortly after, confiding to me in the end that she really didn't see what the big deal was with it and that she would probably never be back.

I understand that some people may not resonate with this outlook -- and maybe it's just me getting older -- but I've grown to see that there's indeed such a thing as going through life in a hurry. I do think that the jury's still out as to the overall impact of AI on what I would label "useful friction".

I imagine how you intended your comment to come across and I get it to some level. But I can't help feeling that there's something a bit dystopian in a world where all friction is removed just to more quickly get to the juicy bits.

I'm unclear as to where your outrage is directed. Is it that they give jobs offshore? Or rather that those who get them are now victim of their original accent not being heard by Canadians?

This is what tends to happen to code when your focus starts to shift away from how expediently you can write it and closer to how readable/maintainable it really is.

If you're a dev, one approach to specialization is to align with the tooling associated with common "profit center" processes. Become a Salesforce/Hubspot/Odoo/Shopify developer. If you're not interested in developing, you can specialize in learning one specific ecosystem really well and then teach companies -- typically SMBs -- how to set themselves up and organize their operations around it.

The article highlights that the real value of the ZK method is in the discovery of the deeper connections that run between ideas that on the surface may appear unrelated.

I can see how that could be useful in contexts where the work is about mulling over concepts, trying to uncover some hidden patterns. Philosophy, sociology, psychology come to mind. But looking at my large cache of notes on well known technology, I have a hard time seeing where the value would be.

I think it's worth pointing out because ZK pops up quite often on HN, as if it's the pinnacle of note taking. In reality, a lot of people here may just be wasting much of their time.

I've written thousands of notes with just vim and the file system for over 20 years with little protocol. It's worked out great for me. Simple short text files that eventually graduated to markdown. I have folders and subfolders for top level topic hierarchy. Usually just a single level, a parent folder and then files for specific topics. It rarely goes deeper than two levels. I title everything descriptively to guide me to find what I need later. Like I said, thousands of notes spanning 20 years, never a problem.

I'm no expert, but looking from afar it seems to me that complex note-taking systems are an optimization on some anticipated theoretical future problem that seldom materializes in practice, and I think trying to squeeze those promised extra 10% of efficiency might possibly qualify as diminishing returns.

saying "what else is new?" is nearly "this does not disgust me" which is essentially condoning it.

That's an amazing stretch.