I am several hundred opossums in a trench coat

  • 1 Post
  • 35 Comments
Joined 2 years ago
cake
Cake day: July 1st, 2023

help-circle


  • I’ve used it most extensively for non-professional projects, where if I wasn’t using this kind of tooling to write tests they would simply not be written. That means no tickets to close either. That said, I am aware that the AI is almost always at best testing for regression (I have had it correctly realise my logic is incorrect and write tests that catch it, but that is by no means reliable) Part of the “hand holding” I mentioned involves making sure it has sufficient coverage of use cases and edge cases, and that what it expects to be the correct is actually correct according to intent.

    I essentially use the AI to generate a variety of scenarios and complementary test data, then further evaluating it’s validity and expanding from there.



  • I think its most useful as an (often wrong) line completer than anything else. It can take in an entire file and just try and figure out the rest of what you are currently writing. Its context window simply isn’t big enough to understand an entire project.

    That and unit tests. Since unit tests are by design isolated, small, and unconcerned with the larger project AI has at least a fighting change of competently producing them. That still takes significant hand holding though.


  • Again, it was a Washington Times article, not Post. The Washington Post was not linked because neither Trump nor Snopes cited them. Likewise, whether or not the “stuff on yahoo” that “seems like ai slop” to you doesn’t change whether it is AI authored (it isn’t, it was written by a human working at Snopes and posted to Snopes) nor whether it is accurate (it is). Trump did post the article with the image in question to his Truth Social account on March 9, 2025.

    The discussion raised by people in this thread is not about the content of the linked Washington Times article, it is about the fact that the president of the United States is using iconography developed by the Nazis in the same manner as the Nazis. That said, to take the obvious bait you’ve set up, we’ve seen how ineffective both Russia and North Korea’s army are. They are clearly a poor model for a well run and organised army, regardless of their supposedly “masculine strength”. I also reject your claim that strength is a purely masculine trait. The US has had a (if begrudgingly) diverse military for as long as it has been a global superpower. Gay people, trans people, people of color, and more recently women have been contributing successfully to that strength for longer than you or I have been alive. Many of those groups are typically cast as non masculine, yet clearly display great strength.

    I’m not going to be responding to you any further, I don’t really feel like you’re engaging in good faith.







  • I briefly worked for a company who worked on household power technology. Their product would attempt to predict energy prices, weather patterns, and usage to sell your excess energy at peak prices. Like discussed in the article, this company collected usage data and controlled the sale of energy back to the grid centrally. They did this because it meant they could better train their prediction models and run them on more powerful hardware. The controllers would have needed internet connectivity anyway to query energy prices, and putting the prediction on device would have just made them more expensive and worse. Even when I worked there (back in 2015 I think), they were already very aware of the threat vectors discussed by this article and took some measures to prevent it.

    In my opinion they were (/are, still exist) a responsible company run by competent people. They did not collect the data out of “greed”, and I strongly suspect that the people in these comments implying that the data is collected to be sold have never actually worked in the industry and have very little idea of the specific value of energy usage data. I can’t really speak authoritatively for other companies, but I would guess that, like the one I worked for, their products are internet connected simply because it improves the product. For example, people expect things to be controllable or viewable from an app from anywhere, and that requires internet connectivity.







  • How much computing power do you think it takes to approximately recognise a predefined word or phrase? They do that locally, on device, and then stream whatever audio follows to more powerful computers in AWS (the cloud). To get ahead of whatever conspiratorial crap you’re about to say next, Alexa devices are not powerful enough to transcribe arbitrary speech.

    Again, to repeat, people smarter than you and me have analysed the network traffic from Alexa devices and independently verified that it is not streaming audio (or transcripts) unless it has heard something close (i.e close enough such that the fairly primative audio processing (which is primitive because it’s cheap, not for conspiracy reasons) recognises it) to the wake word. I have also observed this, albeit with less rigorous methodology. You can check this yourself, why don’t you do that and verify for yourself whether this conspiracy holds up?