Do you really think that if a Chinese model had achieved a significant math breakthrough that it wouldn't be trumpeted to the global media? The "Deepseek moment" was great for China; this would be the same.
Sure, but you lose the representation of the developer’s intention that way. I would be pretty pissed off if I inherited a project and the schema was all ANYs.
The developer's intention is that anything can go in there.
You would only inherit a project where everything was ANY if anything could go anywhere.
With SQLite's default behavior, anything can always go anywhere, so the type definitions are at best semi-accidentally observed by the code, and at worst completely misleading. You have no idea which of the two the developer intended.
I get the impression that this SQLite behavior is a historical oddity caused by the original use case for the tool, rather than something that was intentionally planned and thought through, and was later retconned to be intentional and benign. To me, it makes no sense, even after reading the explanation on sqlite's website.
IIRC SQLite originated as a Tcl extension. In Tcl at the user level "everything is a string" or a number. So it's logical SQLite would accept values as a string or number. Interestingly a Tcl function defines its own semantics, an input value means whatever the function says it means, perhaps a timestamp. SQLite inherited these attributes, and as many commenters observe, SQLite largely continues to work that way.
Implies documentation is crucial. Fortunately SQLite's documentation is among the best out there.
Sure it is. If you encounter something that’s not an int, that could be a signal you have a bug in your writers. Or in the source of the data. That’s useful information compared to “oh, I have some ints and some strings, that’s ANY, everything is ok.”
Exactly. I see software engineering going the way of accounting or lawyer.
Every business needs an accountant and a lawyer on hand. In the past, hiring one software engineer to build custom software for your small or midsized business was not worth it. What can one software engineer build? Maybe an MVP in a year? No chance it was worth it for the vast majority of businesses. Outside of corporations or tech companies, employing a software dev was simply not a thing.
Nowadays, your kindergarten might employ a full time or part time software engineer to build custom software. One dev can build a lot more a lot faster.
That said, I think the average or below average dev won’t earn $200k/year anymore. However, the top devs will earn more than ever. If AI increases an average dev’s productivity by 10x, then it will increase a top tier dev’s by 100x.
It’s already happening. Top tier devs have two ways of earning way more than before:
1. They can build and sell their own products or services. We are already seeing 1-2 person software companies earning real money. Top devs don’t have to stay in corporate if they think they can generate more revenue on their own.
2. When companies get rid of their B devs without losing productivity, they can pay their A devs more.
I’m in the #2 camp right now. My team shrank by 50% through attrition in the last year. We didn’t hire anyone new when someone left. My pay has increased.
It's even better than accounting or lawyers, because good software engineers can build incredible businesses from scratch instead of being tied to the number of businesses that exist
SWEs are more leveraged than ever and we've seen comp drastically rise for top performers
Because there is so much money for kindergarten teachers, paying an extra salary for a software dev for every kindergarten will certainly lead to better outcomes. The computer programs will make up for the lost teacher economically by teaching the children instead of people, raising the market share of my local kindergarten, or enticing people to have more babies. \s
Love sourcehut and want to see them succeed, but their build service (despite having some very cool ideas like allowing you to SSH into your build container) is pretty barebones / lacking compared to GH/GitHub actions. You either get no task parallelism (all your tasks are in one manifest) or you get up to N=4 parallelism (you have four manifests). As far as I can tell, you can’t specify job dependencies beyond just “when this job finishes, trigger this next job by deploying a manifest”. No build caching, and artifact sharing felt like a kludge.
OTOH, the nice thing about a minimalist build system is it forces you to solve these problems yourself, in a way that’s broadly compatible with any provider: for me, nix builds, cache with cachix, and use gnu parallel for running concurrent jobs.
Means that my CI pipeline can be ~instantly moved to any provider, or a box that I own, or whatever.
Language models need redundancy (as informing structure). Not surprising, since they're trained on human language. It's hard to train a model on a language with a high entropy. I haven't tried it, but I think LLMs would perform quite badly on languages such as APL, where structure and meaning are closely intertwined.
I guess that makes sense in the light of her previous post and work on making a new archiving solution for being able to host singlefile archives more efficiently.
reply