Pinned
Links and notes for my talk at #AtmosphereConf about did:plc, how it could break, how we might be able to recover if it did and so forth edochan.com/talks/did/ 5pm today (Sunday), room 2301 Will also mention did:cow if I have time cow.watch
Edmund Edgar
@goat.navy
Free software developer. Live in Mashiko, Japan. Made https://reality.eth.link Mainly post about goats
2.1K followers3.2K following2.9K posts
Pinned
Links and notes for my talk at #AtmosphereConf about did:plc, how it could break, how we might be able to recover if it did and so forth edochan.com/talks/did/ 5pm today (Sunday), room 2301 Will also mention did:cow if I have time cow.watch
Edmund Edgar reposted
Replying to @goat.navy
I used m/🦋’/0 (m/129419’/0) edouard.paris/articles/har...
Is anyone already doing a thing that makes did:keys from a seed phrase using a BIP39 derivation? If so what's the derivation path you use? Speak now or me or the ATProto did:key derivation path will forever be m/231'/0'/0'/0/0, me and @iame.li just decided
This is transitional imho. It's possible to make software correctly, it's just that it's not something humans can do for programs longer than 100 lines or so
Not an ai apocalypse doomer or whatever, but entering a world where the normal operation of cybersecurity is vast swarms of hacker bots battling vast swarms of defender bots 24/7 at a scale beyond human comprehension doesn't seem like a *great* path forward.
Edmund Edgar reposted
That said in constitutional monarchies I think emergency cybersecurity comms should be vested in the Royal Family, give the kings and emperors something useful to do with themselves
Another very based tweet by the Japanese PM. Cuts through the bureaucracy sysadmins have to deal with by telling them, "Client not answering the phone? Middle-management still dithering? Fuck 'em, just shut the website down, if they complain you can tell them I told you to do it"
Also impressed that she replaced her own personal website with the default Apache "It works!", giving people cover to take their sites down even if they don't have time to make some fancy-pants static html maintenance page
Really helpful that she did this (albeit a few hours earlier would have been nice), teams out there are definitely getting yelled at for unplanned downtime rn and a tweet from the PM basically absolves them of responsibility or makes them look good for being ahead of the game
Edmund Edgar reposted
🚨 Six WebPKI Intermediate CAs are compromised. This is bigger than Heartbleed, but a client-side issue. Anything that verifies certificates against public roots, like most TLS clients, is/was affected. Details, live test, patched versions and workarounds at emudhra.we-have-a-problem.com Act now.
Edmund Edgar reposted
This is actually a good idea tbh. Have bots that scan the web and invalidate keys if they're found. Services should mandate you agree whitehats are allowed to do this in ToS so it's not a violation of the CFAA. Would probably reduce their fraud dispute rate considerably.
Oh no, every API key system should have endpoint "invalidate_api_key" If key is leaked, at leat whitehat hackers (or the party who stored them) could invalidate found keys.
Somewhat optimistic about the Farcaster fork by Cassie Heart. The only scenario where a whole community ever moves over at once (not just crumbling off to different things) is "project bought by arrogant comedy villain" which fits Balajis even better than Musk Her emergency telegram t.co/dmpGdZJUT5
Edmund Edgar reposted
Edmund Edgar reposted
Excellent question.
Border collies are made for herding, they enjoy herding and become distressed if they can't engage in herding-like activities. It is unethical not to let them herd or to at least simulate it. Arguably, the mere existence of border collies is unethical, but they exist now, so now what?
Edmund Edgar reposted
The Call for Presentations (CFP) is now open for @atmosphereconf.org We're looking forward to seeing your great submissions!
Edmund Edgar reposted
I think that one of the most successful plays that the tech lobby has made has been to make the tech space extremely unattractive to anyone interested in actual change.
Virkkunen ends her speech, mostly about how the EU will be an “AI Continent” with a reference to Sean’s Heaney’s Digging, saying every generation has its instrument - and maybe our pen / shovel is AI Our culture is really being mined here today Her speech ends and she and An Taoiseach flee
Edmund Edgar reposted
Edmund Edgar reposted
My blog post that explains the latest #WSocial controversy is up: 🔗: blog.elenarossini.com/public-by-de... Disclaimer for W folks: this work is entirely self-funded, with childcare help courtesy of my family so I could write the first full draft on Sunday 💁♀️ #WIdentity #WSocialFacts #privacy
Edmund Edgar reposted
Edmund Edgar reposted
all software is practically open-source now, i can just mod literally any program running on my computer however i want including those with heavy obfuscation and DRM
this is uhm... very good so good that i now regret asking kira to try it out in the first place because it's kind of scary how good she is with it i will not elaborate
Edmund Edgar reposted
Replying to @chancethelawyer.bsky.social
No, it is proving that a bunch of things we assumed were optimal because they “looked nice” in a certain sense were wrong because a far uglier solution was better. Example: the tightest way to pack 9 squares into a big one is with a 3x3 grid of side length 3. The best way to pack 11 is this:
Edmund Edgar reposted
we found the reverse Gish Gallop where Trump-like people just try to lie faster than you can correct them, an LLM can correct lies faster than you can lie
There's growing experimental evidence¹˒²˒³˒⁴ that LLMs are unusually good at talking people out of entrenched false beliefs and conspiracy theories. ¹ pubmed.ncbi.nlm.nih.gov/39264999/ ² pubmed.ncbi.nlm.nih.gov/41179846/ ³ sciety.org/articles/act... ⁴ connecticut.adl.org/resources/re...
I feel like a policy that would solve a lot of abuse problems on bsky would be to ban anyone who is rude to Aaron. The hammer emoji types think they're in charge of the website. We don't need a better way to hide their stuff, we needs a practical statement that they are not in charge of the website
In the interests of transparency, since people will create their own narratives of persecution around the auto-blocker. Here is a sample of the data that I ran an AI query just to look for patterns of abuse across the site, the following is the language that will result in an auto-block. 1/4
Toxic combination of 1. Bsky community is defined in opposition to Elon Musk 2. Elon Musk positions himself at the front of any oncoming trend ...dooms bsky to be reactionary
The thing that worries me most lately is that many people who know that “fascism is bad, actually” have also convinced themselves that AI tools must be eschewed as fake or impure. In a self-reinforcing loop, there is relatively little AI tooling and infra by “fascism is bad” folks.
Edmund Edgar reposted
Replying to @filippo.abyssdomain.expert
“Tech is inherently political” is not just about ideology, but about power, too. Yielding technological improvements to the adversary will have predictable outcomes. I’m really worried it will lock in power centralization in a way that will be impossibly hard to reverse. It’s already happening.
Edmund Edgar reposted
Okay so I've built a new Atlas, this time it lets you explore the conversations on Bluesky over the past 7 days graphically. It should update every ~6 hours and help you navigate the information environment that exists on here. I've learned a LOT about what people do here today... atlas.jazco.dev
Edmund Edgar reposted
Edmund Edgar reposted
Edmund Edgar reposted