Yup. My main living data is on a NAS, my on-site backup is an external USB enclosure with a SATA hard drive hooked up to a Mac Mini, all so that I can leverage a Backblaze account to back up the external hard drive with unlimited data.
Wait can tokens just be directly compared like that? My impression is that token cost can vary by 2 orders of magnitude, depending on model, because the actual work of computation varies by that much between models.
She also did this while she was still a full time undergraduate university student. Normally you’d expect someone with more experience to be doing this kind of low level hardware hacking, but she did all this while still young, before even graduating from university.
Reportedly, the app uses an on-device machine learning model to classify the content.
Using your shiny new phone to train AI…
I would think the models are already trained, and then deployed to phones to do on-device inference. The models should be fixed and static by the time they end up on a phone (in large part because phone hardware is just not technically suitable for training, whether pre training or post training).
True, although that’s kinda on the parents for naming their kid after a character whose arc isn’t finished in popular culture.
Imagine being named Alexa, born in 1998, minding your own damn business when Amazon just completely wrecks your name as you’re entering high school.
Yes, the rules of Conway’s Game of Life is Turing Complete.
The glider shape (the thing moving diagonally out of the window in this meme) can be created non-destructively by a glider gun (another shape that creates gliders), which can be controlled using stored pixels in another configuration within the glider gun, including taking in input from other gliders, such that the whole thing can loop and do everything necessary to be a universal Turing machine. Not super efficiently, but it will still do the things that Turing machines can do, given enough cycles.
The -cel suffix suggests that it’s an insult (derived from incel).
Vibe coding is the name for creating computer code by telling an AI to generate the code for you, without necessarily even understanding anything about the code itself.
A thousand yard stare is a description for someone who looks like they’re staring off into the distance, as if they’re having some kind of PTSD flashback, or just at a loss for words.
This is a photograph of Sam Altman, the CEO of OpenAI, a major generative AI company.
Tokens are the unit of measure of how much computing power a particular generative AI query uses.
So basically, it’s a meme that ironically takes the position of the vibe coder, mocking someone who is actually unwilling to use the AI code generation tools (calling them a codecel and criticizing them for saying something anti-vibe-coding or vibephobic), by giving the blank stare of Sam Altman’s soulless eyes, and calling it a billion token stare, but basically doing it ironically to make fun of generative AI enthusiasts by lobbing a really stupid criticism and making the critic look bad.
The mechanism happens in the way LLMs generate text sequentially. So as the LLM generates text, each time the LLM encounters a choice between two or more equally probable words and has to pick one, it chooses from a known pseudorandom key instead of its previous practice of actually relying on a source of randomness.
If you run the text through a checker script that tries to predict the next word, and takes note of every choice made in any given fork in the probabilistic flowchart, you can see that whole sequence of choices and compare to whether it matches a known key.
You’d need to know the model itself well enough to know when the text reaches a particular fork, and what the choices are in that fork, and the key itself, so even OpenAI/Google/Anthropic may not be able to run this watermark detection on their competitors’ models. But they’d be able to know it themselves, for their own models.
But it’s not the company itself putting those assets up for sale. It’s the court-approved or court-appointed bankruptcy trustee of that company, executing transactions approved by the court. That’s why bankruptcy has to go through court, but it also provides a level of immunity (commonly called “exculpation” in bankruptcy cases) to the individuals actually doing the things ordered by the court.
If it’s illegal to do, a court shouldn’t be ordering it, so the affected people should have to fight the court order before it happens. But it’s not illegal to sell anonymized user data like this, and not illegal to sell employee work-related data, either, in the U.S.
It needs to be a system and a contractual framework that complies with the medical privacy regulations of the place where it’s being used (in the US, HIPAA). I have some doubts about whether specific companies or people will actually comply, or the enforcement of those laws by certain governmental authorities, but I don’t think it’s conceptually impossible to have this kind of tech be compliant with those types of regulations.
The ideas described in Clay Shirky’s Here Comes Everybody and Chris Anderson’s The Long Tail made the mid-2000s super optimistic about having fewer gatekeepers between those who would create something and those who would enjoy that thing: you no longer needed to convince an agent and publisher and distributor for the funds or approval to publish a book, or release an album, or even distribute a short film or movie, because you could just do those things and see if you could find an audience.
And it worked, because on the internet it became economically feasible to publish for a small audience, and it became economically feasible to aggregate a lot of small creators and small audiences into big content-agnostic platforms for the thing being distributed. And a few of them would make it big, too.
Yes, app stores still funnel the overwhelming majority of the money on the platform to a small handful of developers. But I’d argue that the recent moves towards consolidation and concentration is because the ecosystems have betrayed the democratic, decentralized, nondiscriminatory/content-neutral ideals that we built the user-created Web 2.0 on, and instead of descended into enshittified monetization of the platforms’ algorithmic control, SEO-like strategies so that slop isn’t filtered out, and our utterly hacked brains that can’t resist clickbait and ragebait and hornybait to where it’s mostly sorting through slop instead of organically finding the niches we enjoy.
And now instead of traditional gatekeepers, we have tech gatekeepers. It was a fun run, though, from 2005-2015.
on average peoples’ computers
Average people don’t have self-administered computers anymore.
People have work laptops administered by their work IT departments, or they have phones and tablets running iOS/Android. Some children have school-administered Chromebooks, and may never own a normal laptop running a traditional desktop OS.
The main people operating self-administered traditional computers are PC gamers (a dying breed during the current price crisis), freelance/independent workers in the fields who still benefit from a mouse and keyboard for productive work and aren’t provided a computer by an employer, and tech nerds.
Raw lichess data shows that 13% of their games involve a pawn promotion too. Maybe the main character is a pawn who gets promoted, and that’s the plot of the game.
I wonder if the statistics are skewed as “one machine, one vote” rather than “one person, one vote.” Because I know a few people (including myself, when I was younger) who would run like multiple Linux machines and start up random VMs to do a few things here and there.
Knowing that the state that submits the image for human review is that the system must automatically (1) identify a car (2) whose license plate can’t be automatically read, the more productive trickery would be to defeat the “identify a car” part where the software doesn’t even know that a car drive within its field of view, and therefore doesn’t know to take a still frame and forward it to a human.
Alternatively, have the software automatically read a license plate incorrectly so that it doesn’t trigger human review but is instead confidently incorrect in what it did log to the records.
However, each incident will still have a date, time and location in the database along with the captured video of the person, which will make it relatively easy to figure out who among the local population is the most likely candidate.
It sounds like this adversarial clothing is designed to prevent the systems from identifying an image as a person at all, so it wouldn’t be logged. They’d need to rerun the patched software on the entirety of whatever archived video stream they happened to have retained, which is far less assured than just the logs and retained records from when the system detects a person.
Microsoft Exchange and the web-based Outlook probably still has a plurality of email volume, because of their dominance in enterprise. Even on the web based consumer space, Apple and Yahoo still have significant market share.
Fairphone is not about the latest tech. They tend to go for well-supported hardware to ensure longterm support.
The USB 3.0 standard was finalized and published in 2008. USB 3.1 specification superseded it in 2013, and then USB 3.2 superseded that in 2017. Each version is fully backward compatible with every standard version from before, so there’s nothing stopping people from releasing a 3.0 controller and certifying it works in the 3.2 mode that is a drop in replacement for 3.0.
In 2011, Motorola split into two companies, each with rights to the Motorola name and logo.
Motorola Mobility took the business lines related to cell phones. It got bought by Google, and then Lenovo.
Motorola Solutions took the business lines related to everything else, including (and especially) radios for law enforcement, fire departments, militaries, and surveillance video tech.