Skip Navigation
InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)OH
OhNoMoreLemmy @lemmy.ml
Posts 0
Comments 171
Harris has support of enough Democratic delegates to become party’s presidential nominee: AP survey
  • This is going to massively depend on which country you live in, but frequently neither.

    Parties can pick who they like, but they often allow politicians and party members to vote as part of internal selection process.

    In the UK only weirdos and political extremists are party members, and the Tory party tends to spend a lot of effort trying to stop their members from having a vote.

    So of the last four prime ministers.

    Sunak didn't have a vote (lost to truss before that).

    Truss won an internal vote.

    Johnson won an internal vote.

    May was uncontested.

    And this is only the internal vote. All of them became prime minister without an election. Generally you vote for a party (some pedant will claim you vote for MPs, but they do what the party says) and then the leader can change while they're in power.

  • There's European countries who can have several different prime ministers in the span of three months
  • It does. The UK sucks too because it doesn't have proportional representation. But that's a whole separate problem.

    What basically makes it suck a little less than the US is each seat only receives tens of thousands of votes, so it's possible for local interests to do well in particular seats and get a little bit more diversity.

  • There's European countries who can have several different prime ministers in the span of three months
  • So the UK is probably the simplest to discuss because it doesn't have a constitution, and this means parliament is sovereign and decides everything by a simple majority vote.

    They can pass laws saying that certain things need a super majority, but then they can just turn round and unpass them as well

    This means that what you think of as the executive, i.e. the prime minister and all his helpers, can be changed by a simple majority, and an election can be called by one. They don't need to happen at the same time. The last parliament had three different prime ministers without an election, and it's common to switch prime ministers well before an election in order to create an incumbent advantage.

  • Elon Musk's X pushes Trump tags on all US users
  • Nah, you're forgetting how fucked the US is.

    Citizen's United mean it's absolutely fine for billionaires to spend as much money as they like supporting geriatric fascists, providing they don't ask the fascists how they'd like the money spent.

    Providing Elon keeps going off half-cocked and making decisions on his own and isn't following instructions from the trump campaign this is all perfectly legal.

    And even if he is acting on orders, how could you ever prove it?

  • Adversarial images on clothing to combat AI facial recognition without covering the face
  • Yeah but they don't use LLMs for this, they'll use some other kind of machine learning mixed in a big pipeline of data processing. It makes it really hard to guess how much work it would take to fix. It might require retraining, might just require an easy patch of the rest of the pipeline.

    My guess is that they're just shitty jumpers and there's nothing to fix anyway.

  • An Algorithm Told Police She Was Safe. Then Her Husband Killed Her.
  • Unfortunately, this is bad statistics.

    The Teslas in self driving mode tend to be used on main roads, and most accidents per mile happen on the small side streets. People are also much safer where Teslas are driven than the these statistics suggest.

  • Microsoft to Copyright Pi, Found to Contain Entire Arial Font
  • Pi is predictable and deterministic.

    Computer programs exist that can tell you what the next digit is. That means it's deterministic, and running the program will give you a prediction for each digit (within the memory constraints of your computer).

    The fact that it's deterministic is exactly why pi is interesting. If it was random it would typically be much easier to prove properties about it's digits.

  • Microsoft to Copyright Pi, Found to Contain Entire Arial Font
  • No. 1011001110001111... (One 1, one 0, two 1s, two zeros....) Doesn't contain repeating patterns. It also doesn't contain any patterns with '2' in it.

    But pi is believed to be normal. https://en.m.wikipedia.org/wiki/Normal_number

    So it should contain all finite patterns an infinite number of times.

  • Why Facebook doesn’t use Git
  • If I had to guess I'd say it's because fundamentally Facebook development is about deploying servers.

    As you move through the main branch, at any commit, you should have something that you can deploy. The moment you split the repo you lose this, and need to worry about keeping multiple repos aligned.

  • Lawyers for ‘Rust’ armourer move to get case dismissed after Baldwin trial collapses
  • You don't need to bribe a judge.

    You need enough money to have a team of lawyers grind through the evidence and find what's been hidden.

    Compare this to having a public defender with limited resources. They basically have to trust the DA's office.

    What's depressing about this is the DA's office is so used to getting away with this shady shit, that they can't do their job properly even when they know they're under a higher level of scrutiny. Think of all the average Joes that have been fucked over by these guys.

    Rich persons justice isn't really about bribing your way out of things. It's about having enough resources that you can force the system to behave, for you, in the way that it's meant to.

    This is instead of the usual process that just steamrolls over every poor bastard that ends up in court.

  • House Republicans Preview the voter suppression policies they’ll enact if Trump wins
  • It also makes the existing voter suppression much more effective.

    "Oh sorry, none of the black districts have enough staff to run all the polling stations. I guess you'll have to travel an hour and then queue outdoors for four hours if you want to vote.

    Yes, it is illegal for anyone to bring you water, we don't want them to influence your vote, do we?"

  • Sardonic Grin
  • I'm not describing binary classification, I'm describing multiclass. "Group classification" isn't really a thing. Yes, your ml system probably guesses what kind of plant it is and then looks up the ediblity of components.

    The problem with this is how they will handle rare plants that aren't in the dataset, or that are in the dataset but with insufficient data to be recognised.

    Because multiclass assumes that it's seen representative data on all possible outputs (e.g. plant types) it will tend to be dangerously confident on plant types it hasn't seen before.

    This is because it can rule out other classes. E.g. if you're trying to classify as rose, tulip, or daisy and you get a bramble, your classifier is likely to be very certain it's a rose because tulips and daisies don't have thorns. So your softmax score is likely to show heavy confidence in rose even though it's actually none of them.

    This is exactly what can go wrong when you try to use the softmax/standard multiclass approach and come across an interesting rare mushroom or wild carrot. You don't want it to guess which type of plant in the database it's most like, even if this guess comes with scores, you want it to say that it genuinely doesn't know and you shouldn't eat it.

  • Sardonic Grin
  • The key issue here is that 'level of certainty' doesn't really mean what you would like it to.

    You get back a number yes, but it can change according to what's visible in the background, the angle that the plants at, how close is it to the camera, and how nice the camera is you're using (professional photographers use expensive cameras and take shots of different things to everyone else).

    Interpreting this score as "how safe is it to eat the plant" is a really bad idea. You will still eat the wrong plant. These scores can lead to very confident random guessing when you show it a plant it's never seen before.

    And no, softmax is a trick for making the scores all sum to one, so you get back a confidence for every possible thing the image could be of.