> He said he often has to put together demos of AI products for the company's board of directors on three-week timelines, even though the products are "a big pile of nonsense."
@markv I mean, yeah I think it’s almost certainly some sort of consumer fraud?
One thing that “AI” fans have in common with cryptocoin boosters is they all seem to believe technology means regular laws don’t apply for some reason 🤷🏻♂️
Infostealers, if they want to get data beyond what's already stored, have to remain undetected and unremoved for substantial periods of time. People don't log into their online bank every day
But the point of recall is to store everything which means the infostealer only needs to remain undetected for long enough to exfiltrate the SQLite database of OCRed screenshots to effectively get a snapshot of your entire life.
@ianbetteridge Dude! US bank account and social security numbers are used for authentication too! You don't need a password or a PIN to fuck up somebody's life if you have access to these numbers and they're frequently displayed in plain text on screens! The US system is built from top to bottom on principles of security through obscurity and people have no option to opt out of it without literally moving to a different country.
@ianbetteridge I did not write UK. I wrote US. The security issues regarding personal data are very different in the US and Canada and are, yeah, broken in ways that we have to account for when we're designing software.
You keep assessing Recall in terms of whether it's safe for yourself not others. The privacy and security advocates are saying that this compromises the privacy and security of large swathes of the public in a variety of countries. They absolutely aren't exaggerating the issues
@ianbetteridge I don't see how the point still stands. People control what they write down. They don't control what gets displayed on the screens of the services they use. It's very different to write down something in a file with an arbitrary name or an email than to store it in a searchable database with metadata.
And I guarantee that many of the "broken" assumptions are MS reacting to security concerns they hadn't thought of themselves and promising it will be fixed before shipping
You are much too trusting of MS, given their atrocious history for security and privacy and you aren't trusting enough of the people who specialise in security and privacy.
@timbray Voted “Yes” before I saw the “I am one” option. I had brain fog and eyesight issues for months after I first got Covid (despite vaccinations). Still occasionally get brain fog but more rarely, thankfully.
“WebAssembly: A promising technology that is quietly being enshitified”
"So we have this WASI thing that is only version 0.2, and that is already way more complex than the alternative solutions. Is there any way that it will succeed?" https://kerkour.com/webassembly-wasi-preview2
Anybody who thought that OpenAI and Sam Altman were trustworthy before the recent drama has already demonstrated a pretty darn high tolerance for bullshit and shenanigans. A CEO positioning a voice chatbot as a celebrity soundalike without permission isn’t even going to blip on their ethical radar
Debating these people—and there are a lot of them—is pointless. You aren’t going to convince them that they shouldn’t trust OpenAI and Altman any more than you’re going to convince an DHH/Basecamp fan or a Tesla/Musk enthusiast.
It will never tell you that your bad idea is bad. Doing even minimal research would force you to at least consider the possibility that this might be a bad idea with long term consequences about the maintainability of your project. ChatGPT will never do that.
It will never tell you that you’ve made a fundamental mistake in your assumptions or design.
Instead it will help you accelerate full-speed in the wrong direction.