We called this on the Spelunking Pod recently: expect not just more hacks, but more L2s censoring transactions and freezing funds. Here’s Arbitrum doing exactly that.
youtu.be/OXpKqoRYV6Y?...
Posts by Ricardo J. Méndez
If anyone needs a little advice on how to set up (and firewall, and drama-proof) a sponsorship contract for standards work, whether at IETF or W3C or both, that's actually kinda my day job and I'm happy to help. It costs less $$ and drama than you'd think to have "ears on the inside"!
another drawback to all the disinfo floating around about AI causing outages over the weekend: it means you're not primed to see the connection between these two events
a picture is, it turns out, worth a lot more than a thousand words
I’m sure there must be an @oglaf.bsky.social or six about it.
Now, that is not a reason to necessarily use it *exclusively* over @signal.org , but it is the one thing @germnetwork.com provides that Signal does not.
The key thing is being able to create per-context (or even per-contact) cards, instead of having to filter everything through a single identity.
If a context is compromised, becomes irrelevant, or you simply want to cut off contact, you can delete it without any global impact.
Honestly, I don’t know how he manages it. It’s either a kink or a crusade — and the line between those two is thin anyway.
the penne opticon
Sensational bit of checkbook diplomacy by Tether, even if it doesn’t seem to help anyone other than the Drift team.
They got $150M in preferred capital structure. DRIFT holders took the price hit, then got shoved into a structural gimp suit.
solanafloor.com/news/tether-...
If there is a hole — in the code or economic structure — expect motivated hackers equipped with models to be pounding on it.
Any protocol lab not doing hybrid red-teaming right now is negligent.
This is an issue I've been contemplating wrt lexicon design and comprehensibility.
Like it's great for data to be in the users custody, but for whom is the schema designed. There are reasons it abs should be system designers and not data custodians
In theory yes, in practice...
Istm if people lead with doing the sync part first, they tend to get complicated-CDRT-brained and then very rarely actually circle back to making sure data at rest is in legible files. 💁
There's examples where both coexist, but afaik they always started w files.
Hachyderm is a large well run Mastodon instance.
This is from early 2025 (so looking back at 2024 costs). About $1600USD/month for infra.
The expensive part, like with all systems, would be paying _people_.
www.mariosangiorgio.com/link/1137932...
For moderation, not only are current models great at detecting intent in a message based on context (racism, violence, etc.), but I was surprised to learn they’re good even at detecting if an image is meant to be offensive.
Both areas are still people-heavy, but improving.
Not quite — there has been an impact, just not as massive as in code generation.
On the infra side, I keep getting reports from former co-workers about how they can use Claude to diagnose production issues or chase Heisenbugs much faster than they could before. That lowers some costs.
I expect the costs will differ by geography (different moderation requirements), and you will mostly only figure them out when you start scaling.
What would be truly impactful is if people internalized that services have a cost and don’t magick themselves into existence.
“Anyone can fork” is technically true but economically misleading — you can’t fork an operating budget.
Code is free. Infra and labor are not.
Too many investors are pack animals, and act as once-bit thrice-shy. Most got bit back when DeSoc was a blockchain narrative — and we had at least two cycles of it.
The composable data angle might be the better narrative, but it needs some war paint.
“Now that you are married, you’ll need something for when you divorce”
“Oh cool! Steak knives!”
“Um. Yes. Yes, steak…”
I’m sure it looks like it for the cows.
A lot of these scan as “haha, well obviously *I* wouldn’t make that mistake”.
Lovely mate. Good for you. How about all the other things it does better than you or I right now?
I have. If I were you, I wouldn’t be complaining.
The kind people at @eth_limo have warned me that there has been an attack on their DNS registrar. So please do not visit vitalik.eth.limo or other eth.limo pages until they confirm that things are back to normal.
You can…
firefly.social/post/ff-eaaf30997ffb4534...
I wouldn’t chalk that one up to @pds.ls but to Bluesky’s app view refusing to update records with changed hashes, to avoid sneak edits.
blacksky.community/profile/did:...
Technically… mostly hosted wherever your account is, given it’s creating ATProto records. If it’s Bluesky PBC hosting your account, then that data is still States-side.
The US concern can be easily solved by migrating to @eurosky.social, though.
Thanks! Yeah, figured it was a legacy issue. One of those “it seemed like a good idea at the time” things.
Yup. But having the org sign back — eg. @eurosky.social confirming @robin.berjon.com — also lets you prototype and play with two-way signaling.
My knee-jerk was “don’t bother with orgs, they are conservative and require years of social validation”, but…
There are enough ATProto-native teams where that could not only be valuable, but provide an extra layer of trust and validation.