MA woman in Texas is looking at six years in prison for activity on a Cash App account she says she never opened.
CSix years? For an account she didn't open? How does that even get to a courtroom?
MThat's the question. TechSpot has the story — the account was apparently used for transactions that triggered fraud charges against her.
MProsecutors offered her a deal: a thousand-dollar fine, a thousand in restitution, three years' probation. She said no. She says she's innocent.
CTurning down three years' probation to fight a felony charge... that takes conviction.
MOr the absolute certainty that you never opened that account. And here's where it gets ugly.
CUgly how?
MThink about what it means. If she's right, somebody, somewhere, opened a Cash App account in her name. She says she had nothing to do with it. But the charges are on her.
CAnd the system's response is... a plea deal. Just pay a fine and admit to something you didn't do, and we'll make this go away.
MThat's what the Hacker News crowd is furious about. One commenter said it's a carrot dangled to protect careers — admitting fault would be too embarrassing.
CI mean, it's not just Cash App. This is every fintech platform. They make sign-up so frictionless, and the identity checks are... optimistic.
MBut let's be fair. Cash App didn't bring charges. A prosecutor did. The platform's verification apparently failed, and then the legal system took over.
CAnd the legal system's solution is to punish the person who says she's the victim of identity theft, not find the person who actually did it.
MWell, that would require an investigation. Plea deals are cheaper.
COh, come on. That's too cynical. Prosecutors go after what they can prove. If the evidence says the account was used from her phone or her IP, they have a case.
MDo they? Because she says she never opened it. And they're offering probation instead of taking her to trial. That says something about their evidence.
COr it says they want to clear a case with minimal resources. That's a problem too — but it's not proof she's innocent.
MNo, but the burden isn't on her to prove she's innocent. It's on them to prove she did it.
CYou're right. And that's the part that bothers me. The existence of a plea deal says they don't want to fight this in open court.
MThe Hacker News thread has another angle. Another commenter just asked, is it possible US prosecutors give so few shits about justice?
CStrong wording. But when the alleged victim of identity theft is the one facing felony charges, it's not an unreasonable question.
MAnd this is a warning for everyone with a fintech app on their phone. Your name, your identity, is the product. And the recovery from fraud can be worse than the fraud.
CI'll say it plainly: fintech apps ask for a lot of trust, and the system that's supposed to protect you just might charge you instead.
MDifferent kind of AI story now. A user on X says they tested whether an AI can 'remote view' their apartment using street and apartment numbers.
CRemote view? Like, psychic AI?
MThat's what they call it. This account — TheShapeShifter — asked the AI to describe their apartment. Gave it the address as a 'target number'.
CAnd the result?
MOne hundred percent fail rate. Every time. Which, I have to say, is exactly what I'd expect.
CBecause it's not a thing. AI doesn't have a 'morphogenic field' connection. It predicts text.
MBut here's the interesting part. People keep trying this. There's a whole subculture that believes AI has some hidden perceptual ability.
CAnd they always get the same answer. Which they then call a government cover-up or whatever.
MActually, the user framed it as a challenge. 'This can't possibly work. I tested it!' So at least they're on the skeptical side.
CGood. Because if my AI starts describing my apartment, I'm unplugging the router and moving.
MMeanwhile, over on Hacker News, someone built a site to compare how AI models answer the same big questions about humans.
CAnd?
MEvery model, and I mean every one, gives the same 'controversial' take. The poster's word, not mine.
CSame take? They're all trained on the same internet. I'd be more surprised if they disagreed wildly.
MThe specific answer isn't shared in the post, just the observation that they all land in the same place. Which tells you something about training data.
CYeah, that the model's 'opinion' is just the most common opinion in the dataset. It's not thinking. It's averaging.
MWhich is why 'AI wisdom' is often just... the wisdom of the crowd, with the sharp edges sanded off.
CThe wisdom of the crowd, minus the wisdom.
MOne more item, quick one. Someone on Hacker News shared a JAMA study on mortality among people with ADHD.
COh, that's a heavy one. What did they find?
MThe post just links the study, no specific numbers pulled out. But it's a JAMA paper, so it's serious. The title alone tells you the conclusion is likely sobering.
CThere's a growing body of research that ADHD isn't just a school problem. Some people point to its effects on driving, health decisions, and risk. Even lifespan.
MSo this kind of research matters.
CThe science is finally catching up to what people have been living with.
MFrom legal horror stories to AI skepticism to medical research... that's the spread tonight on Tech Tonight.