@miles.land
Doing things at Anthropic. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PAR https://miles.land
Anthropic just dropped new data on how folks are using Claude for emotional support.
Bots can be comforting. But they weren’t built for therapy. Thx to @miles.land & Alex Sanderford from @anthropic.com & @erictopol.bsky.social for insights. In @axios.com AI+.
www.axios.com/2025/06/26/a...
Trump reportedly plans to extend TikTok’s divest-or-ban deadline—again
Congress & SCOTUS backed the law blocking the app from Apple & Google stores
He’s once again showing contempt for checks, balances, & the rule of law
i simultaneously pity the creature and see myself in him
SO excited to share Clio with the world (and on Bsky before Twitter)!
Clio generates insights on AI usage patterns, in a way that keeps user data private. It has unlocked, and will continue to unlock, an immense amount of understanding about the present and future of AI use.
(Blog linked below)
read the paper — there are some fun anecdotes! www.anthropic.com/research/clio
oh
One central concern: It's unclear what the trust anchor is here. The page doesn't say "Miles attests that he is in Mumbai" — it says "here's an authentic photo taken in Mumbai."
Authentic photo from Mumbai according to who? God? Hardware? BBC's visual verification team?
Maybe I'm not understanding the ClickApp threat model. But is it responsible to call this an "authentic photo" taken at a particular location when that location is very easy to spoof? Feels like an exploitable shortcut to public trust.
An "authentic photo" of me in Mumbai. I was not in Mumbai.
There are lots of photo authentication systems, some more robust than others. Anyway, here is an "authentic photo" of me in Mumbai. (Don't trust everything you read from CLLocationManager!) clickapp.com/4hTRfVgqHsw8...
:) thanks!
What content did platforms label as misleading in the run up to the 2020 election? And were platforms consistent in their labeling decisions?
In our new paper, @shelbygrossman.bsky.social, Sam Bradshaw, and I leverage a unique dataset to answer these questions: journals.plos.org/plosone/arti...
I wrote about that time my classmates threatened me (and some friends) with felony charges. miles.land/posts/classm...
a man holding up a sign that says "follow me for the latest AI news"
half the posts in my twitter feed right now
i have never debated anyone due to the fact that it's nerd shit
“You’re not an Amazon driver — you just drive an Amazon truck, work in Amazon facilities, and deliver Amazon packages.”
Ah yes, of course.
The context is that I want to store backup keys in a different location, and I don't want to have to go get those backup keys every time I enroll my hardware security key for a new service.
Is it possible to enroll a hardware security key "remotely" (i.e., without the key itself)? Perhaps using some kind of public key?
I'm not very familiar with the FIDO/U2F/etc specs, and a quick Google search didn't yield anything.
Really enjoyed sharing Atlos.org at RightsCon last week! Working on Atlos — and deploying it to Bellingcat — is definitely a top life experience for me.
chatgpt has bad email taste
imagine sending this email unironically
June 8th, you can see the Global Authentication Project‘s Hannah Bagdasar and Charlotte Godart in an exclusive online only workshop with Atlos outlining how you can design safer visual investigations at scale. This workshop is capped at 50 participants, so get your tickets now!
“Welcome to the emergency room”
…
“Please sign the terms and conditions”
every day, they choose to punish me
meanwhile on the bird site: a bug that lets you like (and bookmark!) a tweet as many times as you want
NAFO activate
spammy follow bots are an important part of any self-respecting social app’s new user onboarding experience