@milesmcbain@fosstodon.org
@milesmcbain@fosstodon.org avatar

milesmcbain

@milesmcbain@fosstodon.org

Code hacker, number cruncher, #rstats user, board gamer, road racer, plant eater, bass slapper. Coming to you from AUS / BNE.
http://micro.blog/milesmcbain

This profile is from a federated server and may be incomplete. Browse more on the original instance.

KathyReid, to microsoft
@KathyReid@aus.social avatar

Why does want to implement ? It's not about images. It's about modelling what workers do on Windows, and then replacing them.

The most expensive part of a computer is the fallible feelings-filled unpredictable meat sack that operates it.

Google has YouTube, Google Photos, Maps, and a bucket load of search data, Google Analytics, advertising, as well as it's data (e.g. transcriptions). And a bunch of data from Android services. From this data they can model speech, model videos and model advertising systems, and how humans respond to them.

But they can't model what people do on computers.

Amazon has Prime data, and a bucket load of compute. But no operating system data. They can build models based around e-commerce and advertising systems.

But they can't model what people do on computers.

Meta has waves hands enough analytics to model human behaviour in the Metaverse.

But they can't model what people do on computers.

Microsoft has GitHub.
Microsoft has LinkedIn.
Microsoft has SharePoint.
Microsoft has Teams.
Microsoft has Dynamics.
Microsoft has O365.
Microsoft has Windows telemetry data.

Microsoft can model what people do on (Windows) computers. Like fill out spreadsheets.Write emails. Synthesize web pages of research. Interact with colleagues on Teams. Create and edit documents.

Microsoft wants data so they can model what people do with operating systems.

Then replace them.

Imagine a CoPilot that doesn't just write buggy code. Imagine one that also does spreadsheets. That creates documents on SharePoint. That communicates with colleages on Teams. That has a customer pipeline on Dynamics.

That's what Recall is about - 360 degree surveillance of the worker, to model their functions, make them fungible, replicable - and replaceable.

milesmcbain,
@milesmcbain@fosstodon.org avatar

@KathyReid at the moment they’re saying everything remains private on local device. So do you believe this is just the start of the enshittification cycle, and they’ll make a grab for data eventually?

jenn, (edited ) to random
@jenn@pixel.kitchen avatar

deleted_by_author

    •
    milesmcbain,
    @milesmcbain@fosstodon.org avatar

    @pbrane @jenn holy shit the microwave is an amazing parallel. I found an old microwave cookbook recently from the 70s, written in peak microwave hype by a person purporting to be a member of the international guild of microwave chefs. The food has the same uncanny valley vibe as generative AI. All these weird combinations of things being made to look like other things for some reason - like carved hotdogs stained with beetroot juice on a bed of warm pinapple.

    milesmcbain, to datascience
    @milesmcbain@fosstodon.org avatar

    With this one the thrust is basically there are a number of "seemed like a good idea at the time" type approaches to reusing data analysis work that deliver benefit in the short term, but will get you absolutely wrecked by complexity and technical debt over the long term. I have found only one scalable way to manage the complexity of building data science capability. Yes it involves writing lots of packages 📦 📦 📦 📦 📦 😅

    https://www.milesmcbain.com/posts/data-analysis-reuse/

    milesmcbain, to rust
    @milesmcbain@fosstodon.org avatar

    One thing I’m appreciating since branching out into and using .table at work is that immutable data by design is an incredibly useful simplification that reduces cognitive load and eliminates large classes of bugs in .

    I can’t recommend packages that try to give users the power of mutability. While the benchmarks are seductive, you don’t gain time because of the new weird bugs and constant struggles over what parts of your API should allow mutability.

    milesmcbain, to cargobike
    @milesmcbain@fosstodon.org avatar

    If you're looking for an electric that can stand-in for a family car in , I've done a long form review of the Tribe Evamos, which doesn't have many reviews of this type online atm.

    https://www.milesmcbain.com/posts/tribe-evamos-longtail-cargo-review/

    milesmcbain, to rust
    @milesmcbain@fosstodon.org avatar

    From the book:

    let mut results = Vec::new();

    for line in contents.lines() {
    if line.contains(query) {
    results.push(line);
    }
    }

    results

    VS:

    contents
    .lines()
    .filter(|line| line.contains(query))
    .collect()

    Guess which goes faster?

    As an dev this result is pretty exciting as it is a total rejection of the 'you have to write a loop eventually' dogma and lets me port my vectorised approaches verbatim.

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Since switching to a rolling linux distro at home package {pak} has taken centre stage. This experiment would be much less fun without it. 🙏 @gaborcsardi

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Just hanging out for the referendum booth data so we can make the highest resolution map yet of Australian racism.

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Unexpected achievement unlocked: handed own code to critique in job interview. Found 2 bugs.

    milesmcbain, to ubuntu
    @milesmcbain@fosstodon.org avatar

    Long time user here, sick of having to hunt down PPAs for recent software versions, so thinking about trying rolling distro with Tumbleweed. Reading very mixed reviews, saying it’s not great for non-sysadmin people. I am reasonably techy but definitely more on the user rather than admin end of spectrum. Am I going to have an okay time?

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    So my holiday reading was the (neo)Vim user manual. It’s very impressive to me how once you get a sense of the model - the concepts and how they relate… this vast surface area of functionality that seemed hard to remember before just becomes a thing you can derive commands in using logic and a handful of higher level principles.

    There are wrinkles for sure, but overall I think it’s an impressive feat of design. 👏👏👏

    I already made myself a minimal IDE in 5 lines of config. 🤣

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Pleased to see the number of cargo bikes on the school run this year has doubled. There’s now more pulling up at school than EVs. Sooner or later now the government is going to get it. 🚲🚲🚲

    TimTeaFan, to random German
    @TimTeaFan@fosstodon.org avatar

    This R rant by @hendrikerz has got quite some attention on the bird site. The R community over here might have missed it due to the missing tag.

    I don‘t agree with much, but some points are worth for discussion / improvement.

    https://www.hendrik-erz.de/post/a-rant

    milesmcbain,
    @milesmcbain@fosstodon.org avatar

    @statsrhian @TimTeaFan chiming in with my 2 cents: r-universe’s search is currently the best keyword / topic package discovery option if you know vaguely what you want. And you’ll find a bunch of gems not on CRAN mixed in: https://r-universe.dev/search/

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Hey context switchers, especially freelancers and consultants:

    I am in dire need of some kind of project task management tool that helps me keep many projects across many clients in the air. Ideally each task has an estimated duration and I can slide them around on a calendar easily, to reshuffle things as clients change their minds etc. Anything you use personally that you would recommend?

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Small thing but I am kind of in love with this {lubridate} method of calculating date differences or durations:

    length_of_stay = start_date %--% end_date / ddays(1)

    start_date %--% current_date >= ddays(90)

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    The CRAN situation in a nutshell is:

    The scope of software being written in and the users writing it has expanded greatly over recent years. The scope of CRAN has not. In fact by virtue of significant policy and process changes it appears to have narrowed.

    For those interested in gatekeeping the community, and status of individuals therin, this appears to be considered a feature not a bug.

    The ‘Comprehensive’ aspect of the name appears increasingly dubious. A great loss.

    milesmcbain, to rust
    @milesmcbain@fosstodon.org avatar

    So 'Rust by example' has all these little runnable examples you can get output for in the browser that's very nice... but what I just realised is that every code snippet in the book is rendered in an embedded text editor!? You can edit the code and re-run with changes to see any behaviour you feel inspired to investigate... 🤯 👏 👏 👏

    https://doc.rust-lang.org/rust-by-example/index.html

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    I’m increasingly convinced Twitter can’t be reincarnated. Twitter was a born of a turn of the millennium internet culture that no longer exists. That cultural lineage was not without problems, but it was Twitter’s value. And in the extermination of that culture Musk has removed all value from Twitter.

    The idea that you can start a new Twitter seeded from Instagram culture? Whatever comes of that it won’t be Twitter. The apple doth not fall far from the tree and all that. 🍎

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    A tale of two programmers:

    1. Believes using a community package is an invitation to break their code, does as much as possible from scratch.

    2. Loves the power and speed of using community packages, but doesn’t use them in ignorance. They pop the hood, read code, debug them, raise issues, make PRs.

    Over time it’s inevitable 2’s taste for software design will mature faster than 1 due to the exposure to different ideas and patterns. They’ll also have more friends.

    milesmcbain, to random
    @milesmcbain@fosstodon.org avatar

    Made this meme for a talk last year and I just successfully cracked myself up stumbling across it again:

    How it feels to be the "data" in data driven decision making.

    milesmcbain, to rust
    @milesmcbain@fosstodon.org avatar

    The CRAN policy doesn't strike me as coherent. "Here's two options one of which we will take away sometime".

    Then they make a strange complaint about crate sources being able to change without their versions changing which is not at all possible for crates submitted to crates.io. This is however possible on CRAN, there are numerous documented cases.

    They also claim building is too hard because crate versions are not fixed. As if cargo.lock files do not exist.

    Drmowinckels, to programming
    @Drmowinckels@fosstodon.org avatar

    📝 New blog post 📝

    'The IDEs I use'

    🧏 People who code have a tendency to spend a lot of time in various IDEs (Integrated Development Environments). They can be as simple as a text editor or as complex as a full-blown development environment. In this post, I'll go through my two go-to IDE's, RStudio and VScode, and why I switch between them rather than sticking to a single one. ---

    👀 Read more at https://drmowinckels.io/blog/2024/ide

    #R #programming #IDE #RStudio #VScode

    milesmcbain,
    @milesmcbain@fosstodon.org avatar

    @Drmowinckels I realised I had all but turned VSCode into vim anyway due to the extensions I am using. And (neo)vim is much more free in terms of how you can create your own tools and automations. For VSCode (and RStudio actually) you’re expected to write and publish a package that contains your extension. It slows things down somewhat. In Vim you just source some code and now things work differently.

    I also have some frustration with VSCode’s design - you can’t avoid a mouse completely.

    milesmcbain,
    @milesmcbain@fosstodon.org avatar

    @Drmowinckels both vim and neovim have a built in terminal now, which I agree is an absolute non negotiable item to use R to its fullest with a REPL.

    This is not a recommendation exactly but if you’re looking to round out your R editor experience you could give emacs + ESS a go. In particular the way they do help is excellent, and inspired me to try to recreate it in RStudio/VSCode with {rmdocs}.

    milesmcbain,
    @milesmcbain@fosstodon.org avatar

    @Drmowinckels I was using emacs for a bunch of different stuff at the time and ESS felt like it offered a lot. Though some complexity is added by the fact they’ve tried to generalise it over SAS, Julia, R etc. If you’re only using emacs for R then yeah I can see RStudio being way more attractive.

    LeafyEricScott, to random
    @LeafyEricScott@fosstodon.org avatar

    What's going on with {plotly}? The R package has 600+ open issues on GitHub, many of them bugs that were reported 5+ years ago and still aren't fixed. Just ran into two of those bugs trying to make a simple line + ribbon plot and only one has a workaround. How are they funded? Is anyone working on these bugs?

    milesmcbain,
    @milesmcbain@fosstodon.org avatar

    @LeafyEricScott the company that paid to develop it went into competition with RStudio/Posit, and the primary R developer went to work at RStudio/Posit. A rift has opened up and the R package has fallen in.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • mdbf
  • ngwrru68w68
  • InstantRegret
  • magazineikmin
  • thenastyranch
  • rosin
  • khanakhh
  • tacticalgear
  • Youngstown
  • slotface
  • Durango
  • kavyap
  • DreamBathrooms
  • provamag3
  • ethstaker
  • GTA5RPClips
  • modclub
  • tester
  • Leos
  • osvaldo12
  • cisconetworking
  • everett
  • cubers
  • normalnudes
  • anitta
  • megavids
  • lostlight
  • All magazines
    •