I think this article is generally insightful, but I don't think the author really knows if they one shotted the excel to python transformation or not. Maybe they elided an extensive testing phase, but otherwise big bugs could be lurking.
Maybe it's not a big deal, or maybe it's a compliance model with severe financial penalties for non-compliance. I just personally don't kind these tradeoffs going implicit.
A thing I ran into is there are no virtual hosts in SSH, i.e. you can't have an SSH server at happy.example.org and ecstatic.example.org if they're both the same IP address. There was a patch for this but support waned. It's a huge blocker.
For any specific use you could build something on top of it to deal with that. For example if you go with the last idea in the article and build http over ssh then your server just looks at the headers.
You would not have separate certs for each subdomain but is that a problem? arguably more convenient.
This list of things not to use AI for is so quaint. There's a story on the front page right now from The Atlantic: "Film students who can no longer sit through films". But why? Aren't they using social media, YouTube, Netflix, etc responsibly? Surely they know the risks, and surely people will be just as responsible with AI, even given the enormous economic and professional pressures to be irresponsible.
> Surely they know the risks, and surely people will be just as responsible with AI
I can't imagine even half of students can understand the short and long term risk of using social media and AI intensively.
At least I couldn't when I was a student.
What is the lesson in the anecdote about film students? To me, it’s that people like the idea of studying film more than they like actually studying film. I fail to see the connection to social media or AI.
It's a funnily relevant parallel you're making, because designing everything around the car has absolutely been one of the biggest catastrophes of 2nd half of the 20th century. Much like "AI" in the past couple years, the personal automobile is a useful tool but making anything and everything subservient towards its use has had catastrophic consequences.
It is political. Designing everything around cars benefits the class of people called "Car Owners". Not so much people who don't have the money or desire to buy a car.
Although, congestion pricing is a good counter-example. On the surface it looks like it is designed to benefit users of public transportation. But turns out it also benefits car-owners, because it reduces traffic jams and lets you get to your destination with your own car faster.
No, it benefits car manufacturers and sellers, and mechanics and gas stations.
Network/snowball effects are not all good. If local businesses close because everybody drives to WalMart to save a buck, now other people around those local businesses also have to buy a car.
I remember a couple of decades ago when some bus companies in the UK were privatized, and they cut out the "unprofitable" feeder routes.
Guess what? More people in cars, and those people didn't just park and take the bus when they got to the main route, either.
But having a car is kind of bad. Maybe you remember when everyone smoked, and there was stuff for smokers everywhere. Sure that made it easier for smokers, but ultimately that wasn't good for them (nor anyone around them).
I am actually, we haven't owned car for years. We also rarely watch TV and eschew social media, so I can still pay attention and analyze things.
But this makes me super weird! This is the whole point of social media bans for kids: if you make it optional, it'll still be prevalent and people making healthy choices will be social weirdos. Healthy paths need to be free and accessible, and things need to be built around them (eg don't assume everyone has a smartphone, etc)
Recently a side discussion came up - people in the Western world are "rediscovering" fermented, and pickled, foods that are still in heavy use in Asian cultures.
Fermentation was a great way to /preserve/ food, but it can be a bit hit and miss. Pickling can be outright dangerous if not done correctly - botulism is a constant risk.
When canning of foods came along it was a massive game changer, many foods became shelf stable for months or years.
Fermentation and pickling was dropped almost universally (in the West).
Laws aren't fictitious just because people/countries break them. No one writes a law thinking "that settles that, no more embezzling." Laws simply tell you how that system works: you embezzle, FBI arrests you, you get tried, etc.
Also the US always made a big deal about not joining various treaties, with their reasoning explicitly being "we actually plan to do a lot of things that would violate that treaty." In that sense, that shows the US actually had respect for those institutions.
Also, the west benefited from this arrangement. Most western countries could benefit from the rules based order, and when they needed a little pump, the US broke some rules and brought home a treat for the home team. You might argue this undermines the whole enterprise, but my counterargument is this is the longest period of relative peace and prosperity humankind has ever experienced, so although it wasn't perfect, it was a huge improvement.
Ofcourse people break laws. But they are enforceable and the authorities have absolute power to enforce them. Putin can get away doing whatever the f he wants but nobody in Canada can get away with breaking any law they want whenever they feel like it, for example. That's the difference between the very real Canadian laws over Canadians and "international law" over nobody. Now Canada can pass a law that is in line with some international agreement, but it's still the law of Canada. Other laws don't apply in Canada. Canadian laws don't apply in other countries. And that's about it. If we had world elections, world government, world police, world courts and world laws, with all countries giving up their sovereignty to those institutions then we'd have "international law". Until then we don't.
International law is different, but everyone knows the scenario where like, the ICJ tries and imprisons Putin is remote. Almost as remote as Trump being tried for treason tho....
I'm not sure "everyone knows" applies here. This is one of these situations where the language is intentionally confusing. Because most people when they hear about laws have certain assumptions about what those are and how they work.
In this case this assumption is completely disconnected from reality. So yes, neither Trump, nor Putin, nor Starmer, nor Macron, nor any US citizen, and likely no citizen, or government of no country with any sort of power (India, China) or with a patron country with power isn't subject to any "international law". I.e. doesn't exist, it's just a word salad to manipulate the masses.
I love Nystrom's writing, and he's so good at it because he's written so much. A huge part of the value of things is how we grow in the making of them, and I worry that in a world where we accept generative slop, we'll never have the opportunity to woodshed enough to become excellent at a craft.
I'm a good engineer because I've written tons of code, I've taken no shortcuts, and I've focused on improving over my many iterations. This has enabled me to be an effective steward of generative coding (etc) models, but will younger engineers ever get the reps necessary to get where I am? Are there other ways to get this knowledge and taste? Does anyone know or care?
We're in the anthropocene now, and while probably everyone who knows what that is understands we have the largest effect on the Earth, it also means we now also have the largest effect on ourselves. We're so, so bad at taking this seriously. We can unleash technology that idiocracies western civilization inside of a generation, I know this because we keep lunging towards it with ever increasing success. But we can't just shamble around and let Darwin awards sort things out. We have nukes and virology labs, not to mention a climate change crisis to deal with. If the US political system falls apart because Americans under 65 spend between 2-3 hours on social media a day, that's a failed state with a lot of firepower to shoot around haphazardly.
And why do we keep building things that enfeeble us? Did we need easier access to delivery food and car rides, or did we need easier access to nutritious food and more walkable neighborhoods? Did we need social media with effectively no protections against propaganda/misinformation? We know that cognitive ability and executive function decline with LLM use. Can it really be that we think we're actually too smart and we need to turn it down a notch?
There are actual problems to solve, and important software to write. Neither algorithmic feeds nor advertising platforms fall under those categories. LLMs are supposed to solve the problem of "not enough software"--Nystrom points at this explicitly with the Washington Department of Ecology ad. But we never had a "not enough software problem", rather we had a "not enough beneficial software" problem, i.e. we'd be in a way better place if our best minds weren't working on getting more eyeballs on more ads or getting the AI to stop undressing kids.
Generative AI isn't empowering us. We don't have people building their own OSes, (real, working) browsers, word processors and spreadsheet programs, their own DAWs or guitar amp modelers, their own Illustrators or Figmas. Instead you have companies squeezing their workers and contractors, while their products enshittify. You can't even run these things without some megacorp's say so, and how are you gonna buy time on the H100 farm when AI took your job?
I'm too tired to write a conclusion. I'm pretty sure we're fucked. But hey look, the cars drive themselves.
> I've written tons of code, I've taken no shortcuts, and I've focused on improving over my many iterations. This has enabled me to be an effective steward of generative coding (etc) models, but will younger engineers ever get the reps necessary to get where I am? Are there other ways to get this knowledge and taste?
I think this will be a problem in the middle term, and I've written about such deskilling before [0]. With the latest crop of foundational coding models and harnesses, and more progress on the way, I'm beginning to wonder if it will matter? If there's a future where agents are designing the code, implementing the code, and reading and reviewing the code... At that point the code is no longer the thing. "Software engineers" will continue to sit at the interface of product and software, but the software will be writing itself. Of course there will be a need for programmers who can actually read and write computer code, the same way there's a need for Fortran and compiler devs today.
The skill that all software engineers will need to learn, regardless of level, is how to leverage commoditized reasoning to build products effectively.
- how to design systems declaratively and in terms of requirements and constraints
- how to configure the systems in such a way that they're automatically testable end-to-end
- how to move tacit knowledge out of people's heads and into the context (all of our meetings will be transcribed; questions from the agent will be generated during the meeting resolve ambiguity; the agent will be an omnipresent attendee in all meetings: "Agent: The topic you're discussing overlaps with what Sally said three days ago when she met with Mike. They covered xyz..."; companies that follow remote work best practices will have an advantage here)
- how to allocate and orchestrate teams of people and agents
> I think this will be a problem in the middle term, and I've written about such deskilling before
I'm thinking more broadly. Here's an example: barring malnutrition, western people aren't as strong as they used to be. But why is that? We have gyms and home exercise machines. People have never had more access to the latest exercise science and technology. It's because you no longer incidentally get the reps in modern life. We're reaping the rewards of that with obesity and cardiovascular disease, and that's bad enough.
Imagine the same thing happening to our minds because we no longer incidentally get cognitive reps in future life. People will be asking chatgpt who to vote for, whether they should have kids, whether they should stay in relationships or which major to choose. People will stop going to doctors because doctors will forget how to doctor after using medical models. Etc.
What's a society that's forgotten how to think like? What happens when there are no teachers, doctors, software engineers, lawyers, writers, artists, therapists, because the lack of economic incentive has made it impossible to justify the employment, let alone the training?
There is value in the doing. We aren't what we produce or make; we are what we do.
> And why do we keep building things that enfeeble us?
To offer a view which is, well, a different flavor of pessimism: The good news is that we aren't enfeebling ourselves, the bad news is many humans are being enfeebled by a smaller group of humans as a form of economic predation.
These problems are all theoretical. If you actually tried to implement them at the scale you'd typically need to sway a federal election you'd find it pretty unworkable. And in close elections, the recount process is pretty intense, so it's even less likely that you'll be successful.
You'll probably want more detail. Ballot harvesting can't work because data analysis shows weird patterns like this ("huh this nursing home went 95% Biden whereas every other nursing home in the county went 55%"). Recounts do signature validation and lawyers from either party can challenge any ballot they want. Voters are contacted to cure their ballots. I've worked on the Democratic side and been heavily involved in doing all of this. We had armies of lawyers, software and data engineers, and organizers.
Most of the pointing out opportunities for fraud comes from a place of like, reasoning from first principles. But elections are huge undertakings involving tons of people. It's hard to successfully commit election fraud at a large enough scale to sway a federal election. It's why foreign adversaries prefer to swarm social media with bots: it has a chance of working.
I'm skeptical of these kind of like, self-describing data models. Like, I generally like at proto--because I like IPFS--but I think the whole "just add a lexicon for your service and bickety bam, clients appear" is a leap too far.
For example, gaze upon dev.ocbwoy3.crack.defs [0] and dev.ocbwoy3.crack.alterego [1]. If you wanted to construct a UI around these, realistically you're gonna need to know wtf you're building (it's a twitter/bluesky clone); there simply isn't enough information in the lexicons to do a good job. And the argument can't be "hey you published a lexicon and now people can assume your data validates", because validation isn't done on write, it's done on read. So like, there really is no difference between this and like, looking up the docs on the data format and building a client. There are no additional guarantees.
Maybe there's an argument for moving towards some kind of standardization, but... do we really need that? Like are we plagued by dozens of slightly incompatible scrobbling data models? Even if we are, isn't this the job of like, an NPM library and not a globally replicated database?
Anyway, I appreciate that, facially, at proto is trying to address lock in. That's not easy, and I like their solution. But I don't think that's anywhere near the biggest problem Twitter had. Just scanning the Bluesky subreddit, there's still problems like too much US politics and too many dick pics. It's good to know that some things just never change I guess.
Not sure I fully get you... In your example, isn't the problem that nobody cares about this data? So there is no motivation to build a client. Whereas if these were beloved notes or minisites or whatever that got wiped out by the latest acquisition (e.g. see https://bento.me/ shutting down), people would know exactly what those are, and there would be incentive for someone to compete for the userbase.
E.g. Blento (https://blento.app/) is atproto Bento that I only saw a couple of days ago. But the cool thing is that if it shuts down, not only someone else can set it up again (it's open source), but they're also gonna be able to render all of the users' existing content. I think that's a meaningful step forward for this use case.
Yes, there's gonna be tons of stuff on the network that's too niche, but then there's no harm in it either. Whereas wherever there is enough interest, someone can step in and provide the code for the data.
I'm saying it would be a lot simpler to just provide an API to download the data. I don't think it's worth the complexity and network of PDS' to preserve this data in perpetuity, but even if I did, nothing would stop me from using the download API and storing tapes in Iron Mountain or whatever.
I think there's also weird cases where you don't really want things replicated: revenge porn, snuff, CSAM, etc. Like, if someone we're like "yes it's true we lose all data forever if we shut down, but we're not indiscriminately pushing some of the worst content to everyone replicating" I'd take that tradeoff immediately.
I downvoted all your recs for polars, 1 because this is a DuckDB thread and it's low-key rude, and 2 because there are 4 of them. I wouldn't have minded if there were a single post that were like "DuckDB is cool, polars could be an alternative if..."
Maybe it's not a big deal, or maybe it's a compliance model with severe financial penalties for non-compliance. I just personally don't kind these tradeoffs going implicit.
reply