The Ai thread

[Katbel]

Amazon sues AI startup over browser’s automated shopping and buying feature

Amazon accuses Perplexity of covertly accessing customer accounts and disguising AI activity as human browsing

www.theguardian.com

 

[dada_dave]

 

[throAU]

With the recent advancements in small models, I'll bet the big guys are shitting their pants.

They're rapidly going to be stuck between "good enough" and "not general AI" in short order.

Samsung recently announced/released a model that beats out some of the massive cloud models and runs in 17M (not billion, million) parameters which will easily fit in most mobile devices.

 

[dada_dave]

 

[dada_dave]

The fun part about this even the AI bros admit the US energy bottleneck is currently causing the biggest hurdle for them as well and claim the lack of such a bottleneck is why China is “winning” the AI race.

 

[Eric]

 

[dada_dave]

As a former scientist this makes me so angry (not at arxiv or biorxiv). Original research can still be submitted for preprint but I worry and this is still a loss for the science community.

Never mind, biorxiv, is all in on AI as an “experiment” replacing human reviewers with AI:



Note that this tool purports to replace peer-review, emphasis on peer. Peer review is far from perfect and academic publishing is a hell all of its own making, but JFC. When confronted as in the thread above they demure that it’s just an option to improve your paper, but in the original blog posts and threads, the claims are sky high.

As someone else noted the most obvious consequence will be AI paper mills running their papers through this tool to get a version that “passes” making it more difficult to tease out AI slop.

 

[dada_dave]

Look it’s easy to anthropomorphize chat bots, which is why some people have so much trouble with them, falling in love, treating them as a therapist or friend, etc … but this is someone, Ryan Grim, who is a supposed lefty journalist, really just a contrarian tankie, arguing in public with a chatbot, trying to threaten it, then backing down because he’s addicted to X and doesn’t want daddy Musk to limit his reach while mocking people who left X. Basically this has layers.

 

[Eric]

Look it’s easy to anthropomorphize chat bots, which is why some people have so much trouble with them, falling in love, treating them as a therapist or friend, etc … but this is someone, Ryan Grim, who is a supposed lefty journalist, really just a contrarian tankie, arguing in public with a chatbot, trying to threaten it, then backing down because he’s addicted to X and doesn’t want daddy Musk to limit his reach while mocking people who left X. Basically this has layers.

Replacing logic with idealism is the MAGA way, now throw all the money in the world at training it in such a way.

 

[Eric]

Essentially everything you see or hear has been ripped from someone who actually created it by AI sucking machines that do nothing but scour the internet for content to repurpose in their own image.