Technology

But EU privacy advocates like NOYB have protested Meta's plan to offer a subscription model instead of consenting to data sharing, calling it a "pay or OK model" that forces Meta users who cannot pay the fee to consent to invasive data sharing they would otherwise decline. In a statement shared with Ars, NOYB chair Max Schrems said that even if Meta reduced its fees to 1.99 euros, it would be forcing consent from 99.9 percent of users.

Element Call is an open source end-to-end encrypted video and voice conferencing solution built on the Matrix protocol for secure communication.

We’re developing Element Call to provide the best possible security properties with compulsory end-to-end encryption, supporting sender verification, forward secrecy, post-compromise secrecy, zero-trust decentralisation and cross-domain capability. It achieves this by building on the foundations of Matrix as a mature, audited [1], open standard protocol for secure communication.

This release updates the Caret Browsing mode to also work in the PDF viewer and adds support for the Screen Wake Lock API.

While the crowd at sxsw2024 booing a sizzle reel of people either promising the beauty of the future “AI” will bring or claiming it to be “without alternative” is funny and went viral for all the right reasons, this event speaks to a deeper shift in perception.

As Brian Merchant writes:

For the buzziest tech of the moment to get shouted down at SXSW speaks volumes about the scale and nature of the animosity generative AI has amassed. The tech is seen, here, as exploitative by tastemakers and by technologists.

But I’d go further: It’s not just the public perception that OpenAI has been trying to plant in our collective understanding is falling apart due to the actions of that strange company, I think the actual narrative of “AI” is untangling.

It’s official — the Pentagon is becoming a bank. Well, sort of. At a March 8th event on dual-use technology at SXSW in Austin, Texas, director of the Office of Strategic Capital Jason Rathje announced that his team has officially received the internal authority to grant executive loans and loan guarantees, a first within the Pentagon.

The Office of Strategic Capital, or OSC, was created in response to growing concern over China’s investment in next-generation technology. According to its investment strategy, released Friday, March 8th, the OSC will invest in firms researching and developing 14 “critical technologies,” including hypersonics, quantum computing, microelectronics, autonomous systems, and artificial intelligence.

There's an extraordinary amount of hype around "AI" right now, perhaps even greater than in past cycles, where we've seen an AI bubble about once per decade. This time, the focus is on generative systems, particularly LLMs and other tools designed to generate plausible outputs that either make people feel like the response is correct, or where the response is sufficient to fill in for domains where correctness doesn't matter.

But we can tell the traditional tech industry (the handful of giant tech companies, along with startups backed by the handful of most powerful venture capital firms) is in the midst of building another "Web3"-style froth bubble because they've again abandoned one of the core values of actual technology-based advancement: reason.

ArtPrompt is what’s known as a jailbreak, a class of AI attack that elicits harmful behaviors from aligned LLMs, such as saying something illegal or unethical. Prompt injection attacks trick an LLM into doing things that aren't necessarily harmful or unethical but override the LLM's original instructions nonetheless.

This is a of a post made during a time where outgoing federation for lemmy.ml was broken. I hope lemmy.ml readers will forgive me for shoving my filthy little words under the shining gaze of their precious and observant eyes for a second time.

I have a Kindle Paperwhite (7th generation). (Stallman weeps) It appears people generally customize their kindle beyond Amazon's original design by jailbreaking it. But I was wondering if I could replace the entire system on the kindle by a new one, for even more hacking fun.

It appears Kindle Paperwhites run on ARM processors, so there should be plenty of compatible software. However, it appears flashing the ROM of kindle only appears in the context of something called the Kindle Fire. Why is that? Is there any reason ROM flashing for the paperwhite kindles isn't common? The only reasons I could think of is that disassembling and reassembling the kindle paperwhite is kinda annoying (especially with the glue holding the case together) and that maybe not everyone has a board to externally flash ROMs. I've also thought that maybe the ROM is write-protected or that the software is signed and that the Kindle will refuse to boot off of anything that hasn't received Jeff's blessing. Is there any existing guide on flashing a custom ROM? Have any ROMs been created already?

Maybe my foolish self has not searched good enough and hasn't found the discussions on ROM flashing of other kindle models, but in any case I think it's good to have this discussion on here on Lemmy too even if it potentially already exists somewhere else on the internet, so that other fools like me may come across your wisdom and be enlightened.

If this is complete and utter nonsense what I'm babbling about, can I at least somehow download the firmware and software running on the kindle from the device, so that I may poke and probe it with my disgusting, dirty little fingers, defiling Amazon's intellectual property?

I hope that you have a good day and that the following days be good too. If I am stupid for even mentioning the idea of a good day, I wish that some day our suffering may end and that a good day be something we all can look forward to.

Audiences attending the SXSW premiere of "The Fall Guy," starring Ryan Gosling, were not happy about having to watch a sizzle reel before the movie that touted the promises of AI.

But now, researchers have devised an attack that deciphers AI assistant responses with surprising accuracy. The technique exploits a side channel present in all of the major AI assistants, with the exception of Google Gemini. It then refines the fairly raw results through large language models specially trained for the task. The result: Someone with a passive adversary-in-the-middle position—meaning an adversary who can monitor the data packets passing between an AI assistant and the user—can infer the specific topic of 55 percent of all captured responses, usually with high word accuracy. The attack can deduce responses with perfect word accuracy 29 percent of the time.