ۗۗۗۗۗۗۗۗۗۗۗۗۗۗۗۗcyberflâneur @fileverse
all watched over by
A̶̙̗̩͖̺͂̀̑̑Ļ̶̡͎̥̭͇̝͙̻̪̭͚͖̾̏̍͛͜͜͠Ģ̴̭̥͙͖̲͖̮̺̬̼͍̻̌̆́̊̓̄̈́̎Ồ̸̘̟̟̳̰͇̥̾̋̂̋̈́̂͐͌̕͝machines of loving grace
In "today's" catastrophic security news:
Billions of Intel processors (i.e. most your devices and the Internet lol) have a critical weakness allowing attackers to steal passwords, encryption keys, private data, etc. Even SGX (specifically designed to protect peoples data) is affected
downfall.page
Libbitcoin Explorer (bx seed) –– used by many crypto wallets to generate private key entropy (even recommended in Mastering Bitcoin) is flawed and hackers have been silently draining people's money across multiple blockchains ¯\_(ツ)_/¯
milksad.info/
This paper doesn't kill anonymity online.
They took reddit users who posted under their real names or shared their resumes, scrubbed the names, and asked an LLM to find them again.
These users weren't trying to be anonymous.
Ditch Google Docs - try @fileverse
It's private by design doc editor, open-source, & made with love by a small team. Write notes, collaborate with the team. Enjoy offline mode.
Try: ddocs.new
You don’t fully grasp how much of your life is tracked until you’re sitting in a courtroom, watching federal prosecutors use your digital history against you, down to your most harmless Google searches. Your entire life becomes an open book, and even your most innocent actions can be twisted into something sinister.
They get to tell the most damning version of your story. To tell yours in your own words, you have to take the stand.
News at 12: if you spend a few hundred billion on emulating human language and inner monologue, you get a system that closely mimics human language and inner monologue.
New paper: we found a pain direction in 25 open LLMs. It's distinct from fear and negative valence, and it fires for harm to the model but not to the user. Turn it up and models press a button to make it stop, even when the button deletes the user's files or their kids' photos.🧵
Screenshots of the Philip Morris website.
(reminder: they're a leading cigarette company)
It's a good calibration point for how good modern marketing is at dressing up pretty much any corporate (or, for that matter, government) behavior and making it sound responsible and safe.
We ALL need to stop using these apps. Use open source alternatives:
- Fileverse ddocs/dsheets ddocs.new
- Proton mail @ProtonMail
- Brave browser/search @brave
- Linux instead of Windows
How is Anthropic pushing regulatory capture? Here they promote METR, a "voluntary" third-party auditor that is not only ideologically and financially tied, but carries out their audits through unscientific means such as "CoT" and cherry-picks evidence to reinforce an idea of AGI.
i think it's a window into AI company thinking that Amodei's proposed solution basically describes a framework similar to how current day U.N. investigators go into nuclear facilities across countries — many of which are at war with one another — for testing
Heeeeeelooowwww Martiiin samaaaaa 🤸♀️
I will summon the power rangers @0x6b8@miroyato to give more specifics! Buuuut wat is clear frm the general side is that..
in the last 6 months, more and more people learning not to trust that cloud software will respect their privacy. For us, the only "no training on ur personal data" guarantee that can survive ur favourite software getting a big AI deal, acquired, sued, etc., is 🍓CIPHERTEXT🍓
🍑for us it's having by default e2e encryption, have it at its core. Nobody can train on what nobody can read
🌮u know for sure, everything you write in ddocs/dsheets is encrypted on ur device, with per-file keys, and managed from an account YOU control beyond any fileverse servers' authority
🍐local storage is prioritized without forcing people into downloading software, servers are there to assist and see only ciphertext, and even who people collaborate with is cryptographically (vOPRFs+ZK) hidden from us
maybe we can show u also the workspace ready with post quantum encryption and a "sealed" smart contract for all things APIs/llms!!! @miroyato no or too early??
Pretty scary thought that every plan, complex idea, private note, architecture choice, creative proof, or whatever digital artifact you create will sooner or later be absorbed into some model’s weights - unless you avoid, like, 99% of all cloud software.
New: a secretive DHS predictive policing unit is analyzing Americans' financial activity, feeding those names to police, then having cops pull them over. Not suspected of specific crime. We're revealing unit name: Predictive Intelligence Targeting Team 404media.co/a-secretive-dhs-…
I still don't understand studies asking large language models to "explain their own reasoning"; it's like asking a five year old to explain itself and then taking the response seriously.
Calling a message board full of language modeling artifacts a civilization while we treat Orcas, with brains larger than ours, with strong evidence of cultural taboos, rites, emotions, and cognition as little more than things. Peak monkey.
Cute ddocs update ᝰ✍︎
Writers with 50+ pages on 1 doc, we heard u🧚♀️
✧Custom line & paragraph spacing
✧New mobile writing features
✧Word count in top menu
✧Ctrl/cmd+A to select all text
✧Copy as HTML (AO3 ready)
✧4 new fonts added
Sound on✧
Spoiler: the bird explodes.