• 0 Posts
  • 217 Comments
Joined 1 year ago
cake
Cake day: June 23rd, 2023

help-circle



  • If you’re somewhere in the world that has a TJ Maxx/TK Maxx or similar, go buy their random products that are on sale. Not all are winners, but if you change up your products and just experiment, you’ll find something you like.

    I have long wavy hair, and right now I’m on a Shea Moisture curl and shine kick, but before then it was the Verb Ghost line of products for a long time.
    Don’t sleep on after shower crap, either. My hair has been really dry lately, so I’ve been using a leave in conditioner by Shea, too (now discontinued, sadly). In the rotation is also the Verb Ghost Oil, and some random peptide leave in. JVN (Johnathon Van Ness) also has some excellent products, but we haven’t found them on sale in awhile.

    I don’t use all the after shower products at once, but each has their use. Once you get a feel for what you’re going for, it’s like having a shelf full of tools.
    And if you got a beard, well… use something and tell me if you figure out what works, because I still can’t figure that out. My hair looks great and my beard looks like it got lost in the desert.




  • (I have another response.)
    Lean away, give her a long look, and gaze deep into her eyes. Tilt your head slightly to the side.

    “…. dad?”

    In the stunned silence, because she never expected you to be right, tell her that you didn’t think you would see her again after she disappeared on her way to get cigarettes all those years ago.

    And then put your hand on her thigh and say “I’ve missed you daddy” with wide eyes.

    Gets ‘em every time.


  • Monument@lemmy.sdf.orgtoGreentext@sh.itjust.worksAnon plays a guessing game
    link
    fedilink
    English
    arrow-up
    26
    arrow-down
    1
    ·
    18 days ago

    “Oooh, I don’t know. Maaaaybeee your lips, because I can’t stop looking at them. But if you did, then your doctor is really good.

    Deliver it light and kind of flirty and she’s not thinking about whatever trap that question is, she’s thinking about how you just told her you want to kiss her.

    If it’s for any other reason - making sure an insecurity is unknown, wanting to springboard into a conversation, wanting to see how you react if she baits you into saying something dumb, or even having her own flirty line to deliver about it - she’ll probably get to it. Assuming you haven’t wooed her into a voracious make out session.


  • I used to play 1v1 Ticket to Ride matches against my wife using the app.

    As background: I’m not a very competitive gamer, but I’m decent at problem solving. When I first learned TtR, I played with fairly … great players. One of my friends was (is?) nationally ranked. They routinely beat the ever-loving crap out of me. I think of the dozens of games we’ve played, I have won maybe 10-20% of the time?

    My wife isn’t bad at TtR, but she doesn’t see things the same way in terms of strategy.

    We had this one game where I drew a bunch of short routes all over the map, which blocked her early in the game, and a series of lucky route draws lead me to connect them, inadvertently blocking her at least twice, including on the last play, where I was just dumping cars to end the game.

    She was always a little upset when I beat her, but this time the discrepancy was so bad and she was so upset. I just stopped playing Ticket to Ride - like, at all.









  • You say “Not even close.” in response to the suggestion that Apple’s research can be used to improve benchmarks for AI performance, but then later say the article talks about how we might need different approaches to achieve reasoning.

    Now, mind you - achieving reasoning can only happen if the model is accurate and works well. And to have a good model, you must have good benchmarks.

    Not to belabor the point, but here’s what the article and study says:

    The article talks at length about the reliance on a standardized set of questions - GSM8K, and how the questions themselves may have made their way into the training data. It notes that modifying the questions dynamically leads to decreases in performance of the tested models, even if the complexity of the problem to be solved has not gone up.

    The third sentence of the paper (Abstract section) says this “While the performance of LLMs on GSM8K has significantly improved in recent years, it remains unclear whether their mathematical reasoning capabilities have genuinely advanced, raising questions about the reliability of the reported metrics.” The rest of the abstract goes on to discuss (paraphrased in layman’s terms) that LLM’s are ‘studying for the test’ and not generally achieving real reasoning capabilities.

    By presenting their methodology - dynamically changing the evaluation criteria to reduce data pollution and require models be capable of eliminating red herrings - the Apple researchers are offering a possible way benchmarking can be improved.
    Which is what the person you replied to stated.

    The commenter is fairly close, it seems.


  • Maybe you can.
    All the money I’d earmarked for kung fu lessons and a collection of random lethal weapons wound up going into pet care and hobbies. Besides, I definitely don’t have plot armor. I’d get popped by some junior security mail cop. They probably wouldn’t even have to shoot me. They’d run me over with their Segway, I’d fall, crack my head open, and they’d put a little skull and crossbones sticker on their scooter, like a WWII fighter pilot.