Matt Stoker

Matt Stoker

@focusreact.bsky.social

Flow Artist, Burner, Software Developer, RPCV

174 followers62 following86 posts

Matt Stoker
@focusreact.bsky.social

News just in, use of ChatGPT makes you dumber, which should surprise nobody whatsoever. Use the technologies at your own risk - the more you use them, the more dependent you're likely to become. These tools are not your friends, and they're not helping you. time.com/7295195/ai-c...

time.comChatGPT's Impact On Our Brains According to an MIT StudyThe study, from MIT Lab scholars, measured the brain activity of subjects writing SAT essays with and without ChatGPT.
Matt Stoker
@focusreact.bsky.social

new LLM jailbreak! -- The method, dubbed “Policy Puppetry,” is a deceptively simple but highly effective form of prompt injection that reframes malicious intent in the language of system configuration, allowing it to circumvent traditional alignment safeguards. www.forbes.com/sites/tonybr...

forbes.comOne Prompt Can Bypass Every Major LLM’s SafeguardsResearchers have discovered a universal prompt injection technique that bypasses safety in all major LLMs, revealing critical flaws in current AI alignment methods.
Matt Stoker
@focusreact.bsky.social

Let us not become dependent on AI, nor build with them systems we don't understand and can't explain. The more we rely on such systems, the more reliant we will become. That future is dark and bleak. futurism.com/young-coders...

futurism.comYoung Coders Are Using AI for Everything, Giving "Blank Stares" When Asked How Programs Actually WorkYoung programmers "can't actually program" because they're too reliant on AI models, writes developer Namanyay Goel.
Matt Stoker
@focusreact.bsky.social

this just in from the AI-vs-content-creators-arms-race department: "It's less like flypaper and more an infinite maze holding a minotaur, except the crawler is the minotaur that cannot get out." #ai #honeypot #infiniteloop

404media.coDeveloper Creates Infinite Maze That Traps AI Training Bots"Nepenthes generates random links that always point back to itself - the crawler downloads those new links. Nepenthes happily just returns more and more lists of links pointing back to itself."
Matt Stoker
@focusreact.bsky.social

Good read, and solid point. We need to separate the concepts of protected "free expression" from "attention manipulation by algorithm, for power or profit." These concepts have blended together on social media, muddying the waters. What would it take to neutralize the algorithms?

The Atlantic@theatlantic.com

Now that the Supreme Court has upheld a TikTok ban, "a window has been opened for acknowledging that, as a matter of law, protecting human expression is qualitatively different from enabling algorithmic manipulation of human attention," Allison Stanger writes.

theatlantic.comThe Difference Between TikTok and Free ExpressionThe algorithmic manipulation of users’ attention is not the same thing as actual human speech.
Matt Stoker
@focusreact.bsky.social

Saying Chinese government control over TikTok doesn't pose an imminent threat to the US is like saying someone holding a gun to your head isn't an imminent threat to you because they have yet pulled the trigger.

Matt Stoker
@focusreact.bsky.social

Something something the law of unintended consequences...

Electronic Frontier Foundation@eff.org

“People looking for alternative social media apps should be cautious about the privacy implications of sharing information with an app that has not yet seen substantial public scrutiny outside of China,” EFF’s @cooperq.com told @theverge.com.

theverge.comAs Americans flock to RedNote, privacy advocates warn about surveillancePrivacy advocates are also warning against using the app.
Matt Stoker
@focusreact.bsky.social

this kind of use of AI is unconscionable. to delegate to an algorithm matters of criminal incarceration, to read its output as Truth, and to go no farther to discern it for oneself - there is no justifiable excuse for it. I'm sure this is just the tip of the iceberg.

The Washington Post@washingtonpost.com

Exclusive: Confident in unproven facial recognition technology, sometimes investigators skip steps; at least eight Americans have been wrongfully arrested.

washingtonpost.comArrested by AI: Police ignore standards after facial recognition matchesConfident in unproven facial recognition technology, sometimes investigators skip steps; at least eight Americans have been wrongfully arrested.
Matt Stoker
@focusreact.bsky.social

it's probably not hyperbole to say that the fate of the Internet hangs in the balance with this case. if the court rules in OpenAI's favor, how are content creators to survive? if content creators can't survive, the entire ecosystem withers. a wise parasite doesn't kill its host.

NPR@npr.org

A group of news organizations is taking ChatGPT maker OpenAI to federal court on Tuesday in a hearing that could determine whether the tech company has to face the publishers in a high-profile copyright infringement trial.

npr.org'The New York Times' takes OpenAI to court. ChatGPT's future could be on the lineIn three consolidated suits, publishers allege that OpenAI broke copyright law by copying millions of articles without permission or payment. OpenAI counters that the fair use doctrine protects them.
Matt Stoker
@focusreact.bsky.social

It seems to me that we need a new image format with baked-in digital signature support, that smartphones can sign as authentic or modified. social media sites should validate those signatures, and communicate that information (or its absence) to viewers. www.wbur.org/onpoint/2025...

wbur.orgDoes a picture prove anything anymore?Millions of people now own smartphones where, with just a tap, you can erase people from pictures -- and even add AI generated content that never existed. What does this mean for our shared reality?
Matt Stoker
@focusreact.bsky.social

of course, this would necessitate storing the signing private key on the device. this won't matter for 99.999% of users who aren't savvy enough to extract it and use it to nefarious ends, but that 0.001% who are savvy enough could wreak havoc. how can we trust what we see in this brave new world?

You are up to date.