Remote
Robin
@esoterra@hachyderm.io
mastodon
4.6.6
Raleigh, NC area technologist, speaker, and writer. I don't use "gen AI" / LLMs in any of my work.
0
Followers
0
Following
48
Posts
Joined February 26, 2023
Posts
Replying to
@esoterra@hachyderm.io
@limeleaf@social.coop to be clear, a co-op run git-pages instance would still be pretty great.
Replying to
@limeleaf@social.coop
@limeleaf@social.coop I definitely thought this was gonna be a colo offering bare metal servers based on the info in this message.
Is this essentially a co-op run git-pages instance or is there more to it?
Replying to
@david_chisnall@infosec.exchange
@david_chisnall@infosec.exchange @whitequark@social.treehouse.systems I think the actor model is a hard fit for embedded. On the other hand, systems based on bounded blocking channels (something something CSP), should work fine. I think TinyGo is a good example of this.
Replying to
@david_chisnall@infosec.exchange
@david_chisnall@infosec.exchange @whitequark@social.treehouse.systems that makes sense. At that point, when some core quality is lost, changing how the system works and can be used, a subtle line is crossed between Erlang variants and Erlang-inspired languages.
I think in a way that encourages us to ask useful questions like what properties of Erlang are you interested in? what are your other inspirations? what requirements do you have? what are the implications of combining those things in this way?
(though obviously you don't have to have answers or be interested in them. I just think it gives useful structure to the inquiry)
Replying to
@b0rk@social.jvns.ca
@b0rk got it! hope you find what you're looking for.
this surely isn't it, but I can't help but mention
https://xkcd.com/color/rgb/
Replying to
@joelle@social.joelle.us
@joelle but if you truly think about it, if everyone is concise, then concise is everyone. whereas verbosity is an ouroboros vomiting its tail in infinite joy. praise to the Word God. word of God? the Word is God? more Words, more god!! free trial does not include god, but in 2 more years it will!!
Replying to
@peterhuene@hachyderm.io
@peterhuene@hachyderm.io yup! like this post argues, each time you fail a test and make adjustments, you use up some of its descriptive power. LLMs do this at *scale* making the tests useless.
https://types.pl/@wren6991/116488366700176901
Replying to
@yosh@toot.yosh.is
@yosh @menos I don't think it's "the one true way to talk about uptime" but "percent of time the system is not degraded" is something people care about.
Let's call them AND-uptime and AVG-uptime. AVG-uptime is a better representation of how the component services doing, but AND-uptime is a better measure of how the system as a whole is doing.
Especially, because many use cases will break if any part of the workflow does. If I'm trying to land a PR (git-ops) that runs actions where some steps use the API and load packages, I'm dependent on like 4+ of these services and the AND-uptime of them matters to me.
Replying to
@esoterra@hachyderm.io
@yosh I'm curious what you think of the historic uptime plot someone made
https://damrnelson.github.io/github-historical-uptime/
Replying to
@esoterra@hachyderm.io
@yosh beyond that, the "real GitHub Uptime tracker" captures something important to people "some shit is broken all the time" even if you don't like that they're calling that measure uptime.
Replying to
@yosh@toot.yosh.is
@yosh it's a subtle thing, but I don't like this claim:
"A downstream effect of this calculation is that your uptime numbers look worse if your services are well-isolated."
if your services are well isolated, meaning an outage in A doesn't cause an outage in B, then that doesn't make you look worse than if they weren't isolated by these metrics, it looks the same.
in fact, that they do look the same is their critique.
Replying to
@aesthr@wandering.shop
@aesthr it was really hard to find lol 😝
https://duckduckgo.com/?t=fpas&q=synthwave+image+generator&ia=web
Replying to
@aesthr@wandering.shop
Replying to
@Craigp@mastodon.social
@Craigp eh, I think it sucks and the name sucks, but I think it started out sucking cause it's a bad design not because the name was bad. it just sounds like the original contributors had bad taste in both.
Replying to
@ryan@social.eqrion.net
@ryan have you taken a look at the reference manual @sunfish wrote?
https://github.com/sunfishcode/wasm-reference-manual/blob/master/WebAssembly.md
Replying to
@ryan@social.eqrion.net
@ryan I would love to contribute to this. When I first got involved in Wasm, the spec was extremely unapproachable, and I had to piece together an understanding from the spec, blog posts, and projects. There's been a need for more docs like this for a long time and I have a lot of thoughts on it.
However, I won't contribute my writing to a project that allows LLMs to be used. To me, it devalues the human time, care, and attention I put in my work. Please add and enforce a no LLMs policy in your contributing rules/guidelines.
Replying to
@ekr@mstdn.social
@ekr@mstdn.social First, I want to acknowledge that Mozilla and Firefox are in an impossible position. Too much is asked of them at all times and as the non-profit-backed alternative to Google and Chromium, Firefox has been volun-told to be the champion of the internet's hopes and dreams for freedom and privacy.
There are many things Mozilla and Firefox cannot do without receiving significantly more criticism than anyone else, and surely more than they deserve. But they are different, they are seen as different, and there are different rules.
For me personally, investments in Rust, Servo, and Wasm are the kinds of things that align with what I want from Mozilla. In many respects, you could say Mozilla succeeded in creating and incubating these projects, and I think it's great that Wasm lives on as an active standard and Rust and Servo have their own independent multi-stakeholder lives.
I don't know what motivated the layoffs or the ins and outs of Mozilla's finances, but pulling back from these efforts and putting some of them in more precarious footing, isn't going to make me feel warm and fuzzy, especially since I know some of those people and care about those projects.
I want to return to my original claim:
"Mozilla lacked either the resources or will to support [Servo]"
This isn't a value judgement. This isn't saying that they made the wrong move. Y'all did something I don't like. I gave two reasons that seem to cover my bases and are pretty fair. I don't know and will never know what I would have done in your shoes. I'm not in them.
Replying to
@nasser@merveilles.town
@nasser neat talk! I wasn't sure where it'd go from the title, but "build your art on a sustainable practice or make that practice more sustainable and identify and respond to the systemic threats" is a pretty good answer.
I was more asking if you have any examples of an HTML file for a presentation though? I read some of the library and I'm curious what it looks like to use it in practice.
Replying to
@ekr@mstdn.social
@ekr@mstdn.social I'm not gonna call it "the dumbest possible decision imaginable" but I am going to say that mozilla has made a lot of dubious moves over the years, and that while they certainly got something out of the Servo project, I doubt it was built to be abandoned and seeing Servo continue as its own project shows the promise it had, but that Mozilla lacked either the resources or will to support.
Replying to
@soc@chaos.social
@soc@chaos.social are you thinking of Servo? it split off and if it's own thing now
Replying to
@neauoire@merveilles.town
@neauoire that sucks. at least the irony of it is a little funny, even if it's also kinda depressing.
Replying to
@dlants@hachyderm.io
@dlants@hachyderm.io interesting, essentially de-sugared Rust. I can see why this might be easier for LLMs to parse and generate, but I'm not convinced it's going to be pleasant to review 😅
also, I recommend this talk of you're interested in pursuing this as a long-term project
https://youtube.com/watch?v=XZ3w_jec1v8
Replying to
@nasser@merveilles.town
@nasser one place I would hope to be an opportunity to address this, and could still be someday, is in the Wasm Component Model and WASI, but for now at least the identifiers in its IDL in Components themselves are restricted to the Latin alphabet. I'm curious what Luke Wagner, @sunfish, and other folks think about broader language support in Wasm. Ideally Wasm Component composition would let us link components across programming and human languages and help address this someday.
Replying to
@ObsidianUrbex@mstdn.social
@ObsidianUrbex ooh, this site has a 360 panorama of it!
https://abandonedin360.com/abandoned-government-properties/kelenfold-power-station/
Replying to
@yosh@toot.yosh.is
@yosh@toot.yosh.is plus, you often get higher efficiency or lower wear and tear running equipment with more margin
Replying to
@michalfita@mastodon.social
@michalfita@mastodon.social there are multiple bit patterns that can represent a NaN value.
sometimes these are used to encode different cases, sometimes only one is used practically.
this is why checking for a number being NaN usually looks like isNaN(n) and not n == NaN.
Replying to
@bjorn3@hachyderm.io
@bjorn3@hachyderm.io @sunfish@hachyderm.io maybe this kind of thing should require help from the runtime though so you know your RNG sources aren't getting messed up or intercepted in basic composition scenarios.
Replying to
@sunfish@hachyderm.io
@sunfish@hachyderm.io I guess at that point you could have an API that returns a seed that's unique per instance, but maybe you might as well just let it return multiple at which point it's just an insecure RNG API and maybe that's fine or even simpler.
Replying to
@sunfish@hachyderm.io
@sunfish@hachyderm.io I know, I was just thinking through it for a specific example and arriving at the conclusion that if embedding PRNG+seed is fine for insecure random numbers, we probably don't need a built-in for them and reproducible/replay-able components probably should never touch the secure built-in.
Replying to
@josh@social.joshtriplett.org
@josh@social.joshtriplett.org @sunfish@hachyderm.io if it's a local/internal concern, do we need to? the flexibility of being able to choose the seed, how exactly you embed it, and the PRNG algorithm you use seems nice. you even get to have multiple of them if you want. plus, it means we don't have to do even more standardization work.
what's the value of standardizing arbitrary non-secure random numbers in reproducible/replay-able components?
Replying to
@josh@social.joshtriplett.org
@josh@social.joshtriplett.org @sunfish@hachyderm.io can't the game just embed a PRNG and seed in itself?
Replying to
@sunfish@hachyderm.io
@sunfish@hachyderm.io to pick a specific use case, how about reproducible builds?
you shouldn't need to do any cryptography that requires secure random numbers. e.g. you might verify signatures, but not generate keys
you might need random number generators for certain algorithms though and those will need to be portably reproducible.
I guess you could always embed the PRNG and seed within the component for those though.
Replying to
@yosh@toot.yosh.is
@yosh@toot.yosh.is @sunfish@hachyderm.io sure, but I imagine we could specify how to handle this in a deterministic profile of the component model. maybe a PRNG with a seed computed from the component's static representation?
Replying to
@ikolikoli@mastodon.social
@ikolikoli@mastodon.social @geospacedman@mastodon.social @Frieke72@mastodon.social it doesn't automatically mean 100% for sure there is other generated content, but it is definitely a decent indicator. I don't want to look at, read, or listen to generated material. I am interested in participating in a human discourse which other people put themselves into. So I'm going to choose to only attend talks that reflect that consistently and not ones where I have to squint at the whole thing and wonder whether there's anything the speaker actually did themselves.
Replying to
@dpfyhrie@fediscience.org
@dpfyhrie@fediscience.org @geospacedman@mastodon.social it's also a matter of respect for your audience imo. creating your own visuals or finding relevant visuals is part of the work of being a presenter, if they abdicate it, what does that say about them? more broadly, if a human can't be bothered to make something, why should I be bothered to read/watch/listen to it? the audience's time is precious and speakers are not entitled to it just because they typed in a prompt.
Replying to
@xyla@shitpost.trade
@xyla@shitpost.trade when I read "attention all units" I hear it in the voice of the Halo Wars announcer saying "all units" when you select units.
Replying to
@ricci@discuss.systems
@ricci@discuss.systems @regehr@mastodon.social ya, this has big "if you think technology can solve your problem, you don't understand technology and you don't understand your problem" energy
Replying to
@diffractie@glitterkitten.co.uk
@diffractie@glitterkitten.co.uk with another billion faces on the dice, I think we'll get it to lie way less!