|
Alan Smithee posted:https://twitter.com/JasonColavito/status/1611710986871767041?s=20&t=cN8alEbHkkLxMELPFfdYgA i'm surprised to learn that google screens ai text given that that's every search results they provide me except reddit, wikipedia and stackoverflow
|
![]() |
|
![]()
|
# ¿ Mar 24, 2023 11:41 |
|
> grob flappler review - Which Grob Flappler should you buy in 2023? - Best Grob Flapplers for 2022 - Grob Flappler Review for iPhone/Android - Grob Flappler vs Flab Grobbler Feature Comparison - The Top Ten Grob Flapplers of 2023 In this article we will look at the best Grob Flapplers you can buy in 2023. Comparing Grob Flapplers can be difficult, but we've done the hard work of comparing them all and comparing their varied features. With so many Grob Flapplers on the market, this article will tell you how to find the best Grob Flappler for your money! But first, what is a Grob Flappler?
|
![]() |
|
infernal machines posted:it's always ffmpeg someone please get me that post about the foss tool with a weird name without which the internet would collapse. although ffmpeg bucks the trend by being maintained by not just one guy in czechia
|
![]() |
|
tyvm
|
![]() |
|
i skimmed it and had a post ready but then i clicked the wrong button on my phone and lost it my vague understanding is that there are 4 NNs building on each other, one of which is trained by mimicking yet another network, so there are lots of places some error can creep in. it doesn't build a dense 3d image directly from wifi, it uses a network trained on wifi and camera to build a 2d "image" from 3 wifi signals, then does keypoint extraction on that "image" to detect humans and joints and stuff, then turns that into 3d. so probably if a keypoint gets missed that entire body part just vanishes. likewise in the second row it seems like it wrongly labelled part of a chair back as a head keypoint, and then hallucinated a whole 3d head model to go with it
|
![]() |
|
it's cool to think that there are likely thousands of companies out there doing similar leaning towers of NNs to decide your insurance or credit score or healthcare, except nobody is making them publish how fragile they are (maybe the EU is going to, a little bit, eventually)
|
![]() |
|
i do think that's a pretty cool idea in the paper though and it's insane that it works even a bit
|
![]() |
|
NoneMoreNegative posted:quite the opposite! im telling my kids this is what a cd walkman was
|
![]() |
|
Jabor posted:you might find this take from a fellow educator interesting: https://acoup.blog/2023/02/17/collections-on-chatgpt/ this guy understands chatgpt and ai better than almost any tech journalist writing about it. the ted chiang piece he references is worth a read though
|
![]() |
|
weird how the guy trained in critically reading sources and investigating unreliable claims in pursuit of putting together a cogent thesis is easily able to see through openai's bullshit. i love how the structure of the piece itself is a rebuttal of the arguments he's debunking - it's a perfect example of the kind of essay chatgpt could never produce
big scary monsters fucked around with this message at 17:08 on Feb 22, 2023 |
![]() |
|
Improbable Lobster posted:i think that a lot of very stupid people got told that computers work like the human brain when they were young, assumed the opposite was also true and now think any moderately convincing chatbot is skynet not really helped by calling our big piles of sums "neural networks" with "neurons" that perform "learning" and "artificial intelligence". it gives people all kinds of stupid ideas
|
![]() |
|
![]()
|
# ¿ Mar 24, 2023 11:41 |
|
i think there are already essays on postmodernism you can go and read if you want
|
![]() |