> Cloud makes sense when elasticity matters; bare metal wins when baseload dominates.
This really is the crux of the matter in my opinion, at least for applications (databases and so on is in my opinion more nuanced). I've only worked at one place where using cloud functions made sense (keeping it somewhat vague here): data ingestion from stations that could be EXTREMELY bursty. Usually we got data from the stations at roughly midnight every day, nothing a regular server couldn't handle, but occasionally a station would come back online after weeks or new stations got connected etc which produced incredible load for a very short amount of time when we fetched, parsed and handled each packet. Instead of queuing things for ages we could instead just horizontally scale it out to handle the pressure.
We've switched over our libraries at $WORK to use `tsdown` and it's mostly been a very boring journey, we switched from `tsup` and the DX gains have been massive. Running our `dev` process in the frontend monorepo compiles and bundles all the libraries in less than a second on a cold start compared to `tsup` which was far slower. The biggest gain however was in our CI/CD pipeline where the build servers are much weaker than our developer machines, the `build` step in the quality gate for example went down by over a minute. We've also switched to the new native `tsgo` [0] for type checking, saving us another minute on CI/CD and have migrated a few things from ESLint to Oxlint, which was another easy minute saved. And we switched from Prettier to Biome, and checking the formatting on CI went from ~15s to ~1s. Massive gains are being had in the JS-world from gradual oxidation. Can't wait for Vite with rolldown, we tried that but have a few libraries that depend on SWC which made it a show stopper.
I used Org Mode for all mandatory assignments [0] when I was a student, it worked super well and I could create well formatted LaTeX documents/PDFs from it as well. The supervisors were very impressed that I did my assignment in LaTeX, and I could never be bothered to correct them :-)
Very much this. In fact, I had made a note about earning "professor points", in my (giant) blog post [0] describing my org-mode use...
6. Professor Points: I also get intermediate tex output for free,
which I don't need, but may please you if you are a professor
(or said professor's hopelessly sleep derived grad students).
---
[0] Discussed here recently:
Why and How I use “Org Mode” for my writing and more (2022) (evalapply.org)
217 points by sebg | 6 months ago | 64 comments
ah ... hahaahhh ... quite possibly, I was sleep, ah, derived at the time of finishing it up ... i shall let the typo remain, to vex yet another eagle-eyed reader
Kudos to you, my biggest regret nowadays was not learning how to use orgmode (and org-roam) in college. As long as you set up a proper environment and have an org-publish config you barely need to bother with verbose LaTeX documents.
I had an amusing incident that was related, quite some years back as a CS undergrad. One prof had quite poor eyesight, so had a no-pencil rule for assignments. Typed was OK, neat pen was fine also. (My writing has always been atrocious, typed it is.) As it happened, I used LaTeX (and I knew she did also.)
She also had a rule about promptness of assignment delivery, meaning the start of class when she arrived. Fair enough!
On this one assignment, I had done all but the last question at home, and printed it off, with the hope that I could finish it at the lab and send it to a laser printer in another building. At the time, labs used shared, dot-matrix printers that were usually jammed. I got the last question done, sent it to the printer, then checked the queue. Apparently someone was printing a book, not going to get done before class.
I emailed her my .tex file, with the explanation that I'd hand in my incomplete printout in class, but would appreciate if she could mark the last section so I knew how I did. This was sent well before class, as the timestamp would show. The prof ended up printing, marking, and returning this version, and I received full marks!
So much negativity... I for one really like the look of this, I've been yearning for something that's not the current crop of frontend frameworks to see where we go. The author definitely has the past experience to build something like this based on being a core member of both Svelte and React. It'll be cool to see where this goes, though I wish we had more Elm-like frameworks (that don't lock down the compiler).
If you've never been exposed to a Hindley-Milner type system[1] it can seem a bit magical, but it essentially works by trying to figure out the types from the inside and out by inferring usage all the way to the top. The type of `n` however is `&str`, but I take it you mean the matching. `n.parse()` can be anything that implements `FromStr`, but `Token::Operand` can only take a `u32`, so it can immediately infer that the result of `n.parse().unwrap()` must be `u32` (`n.parse()` is a `Result<u32, Err>`).
I think an EU membership will only happen in Norway when the old generations die out, the yes/no divide is a big generational thing. Even though the latest poll only shows a six percent difference between yes/no, it's only as high as it is now because of the current political situation with trade wars and actual wars. Any vote now for a membership will only happen when a party knows the yes side will win, doing otherwise will postpone the vote another thirty years again. Personally I think we should join the EU, we're already heavily tangled with EU laws and regulations, but I understand the concerns for our agriculture and fishing industries.
What would make your agriculture and fishing industries less competitive by entering our market? I understand EU gives massive subsidies for those, wouldnt your farmers also benefit from those? EU money is distributed regarding the land usage [0], not the production quantity or sustainability. Norway seems a big country and the big deers (or whatever) ranchers would thrive with our system as well as does the biggest ranches in other EU countries. Don’t get me work, this system is unsustainable as it promotes the most land intensive food creation and EU still rely on worldwide importation for its animal feed needs.
Norway has incredibly high tolls and tariffs on importing food (a bit ironic nowadays), primarily to protect the local farmers etc. We're a big country, but the amount of arable land is small and the climate does us no favors. We're at about 39% self sufficient for food, and import a majority of our vegetables, grains and fruits (Norway is one of the biggest importers of food in the world). Fact of the matter is that the only thing that props up local farmers is the subsidies and high import fees, if we join the EU that will have to go and local farms will not be able to compete on the open market. There's obviously much more nuance and details here, but that's the gist of why a lot of people don't want to join the EU. The last vote was essentially decided by the agricultural industry in Norway.
I recently bought an old Brother knitting machine (KH940) which is electronic but with a community of hackers adding third-party firmware and hardware on it [0]. There are also lots of models that read punch cards [1] and knit that purely mechanically (and later models electronically). They are a marvel of engineering that has essentially died out, only Silver Reed and Taitexma produce new models as far as I'm aware and they are often not as featureful as the old machines (SR has no garter carriages that Brother machines had 40 years ago for example).
A while back, Lee Valley had a 3D knitted chisel roll available for purchase, which I've always regretted not buying --- any idea on who might have made that or where to get now?
There are commercial machines, but it's important to understand that in the textile industry there's a big difference between commercial and domestic machines, and it's not necessarily that commercial machines are better.
Besides the enormous price difference, commercial machines are usually less flexible than domestic machines. They're optimized to perform a single task at production volume, rather than to do a variety of tasks. Much like how commercial sewing machines typically perform only a single stitch (as opposed to domestic machines, of which modern examples can perform an arbitrary number under computer control), commercial knitting machines can usually only perform a single type of knit. This does mean that the decline of domestic machines in some parts of the textile industry leads to a loss of capabilities that used to be available.
There are industrial knitting machines, but they’re orders of magnitude more expensive, heavier, and tied to closed software ecosystems. They’re extremely capable, but they’re also far outside the hobbyist realm. The only real evolution on these old hobbyist machines is the Kniterate at $16k and 600lbs.
I’m a knitter who’s been considering doing something like this for some time. Thank you for these links. That second one seems like a great jumping off point for this whole world.
A shredding company that automatically scans and uploads your documents before securely disposing of them. The scanned text is then used to train AI. It’s like the physical equivalent of ignoring robots.txt.
I also over-engineered how I generate my CV[1], but went the opposite direction by using Dhall to create JSON and LaTeX files that I use to create a PDF and GraphQL API in Rust for it, automatically deployed via CI/CD to a VPS and a tagged GitHub release. It was a lot of fun to make, but is so over-engineered I hardly want to touch it anymore :)
I too would be immensely frustrated if this was released in my terminal emulator, and would jump ship immediately. My current job would (and probably will) ban this and similar integrations unless compliance gives the thumbs up... and they do not care about how optional it is. GitHub Copilot is still not allowed unless you are in a small pilot project with their own agreement and contract for sharing and storage with Microsoft with tons of red tape. ChatGPT is still banned. Please keep AI out of the core features of applications that absolutely don't need it.
This really is the crux of the matter in my opinion, at least for applications (databases and so on is in my opinion more nuanced). I've only worked at one place where using cloud functions made sense (keeping it somewhat vague here): data ingestion from stations that could be EXTREMELY bursty. Usually we got data from the stations at roughly midnight every day, nothing a regular server couldn't handle, but occasionally a station would come back online after weeks or new stations got connected etc which produced incredible load for a very short amount of time when we fetched, parsed and handled each packet. Instead of queuing things for ages we could instead just horizontally scale it out to handle the pressure.