jonny (good kind)
Matt Blaze
kottke.orgNothing Works in Trump’s America — Except Racism. “Trump is objectively bad at running the government, but he’s objectively good at running a Klan rally, and his supporters value the latter so much that they forgive the former.” https://www.thenation.com/article/politics/newsletter-laguardia-tsa-baseball/
@curtosis
- 265 Followers
- 163 Following
- 6.9K Posts
Computational linguist, artificer, λ(data). Syntactician with semanticist delusions. Theatre artist. Anti-Taylorist. Continuous everywhere, differentiable nowhere. Boosts are multivalent. He/they.
reminder that anthropic ran (and is still running) an ENTIRE AD CAMPAIGN around "Claude code is written with claude code" and after the source was leaked that has got to be the funniest self-own in the history of advertising because OH BOY IT SHOWS.
it's hard to get across in microblogging format just how big of a dumpster fire this thing is, because what it "looks like" is "everything is done a dozen times in a dozen different ways, and everything is just sort of jammed in anywhere. to the degree there is any kind of coherent structure like 'tools' and 'agents' and whatnot, it's entirely undercut by how the entire rest of the code might have written in some special condition that completely changes how any such thing might work." I have read a lot of unrefined, straight from the LLM code, and Claude code is a masterclass in exactly what you get when you do that - an incomprehensible mess.
So in summary, mail-in ballots are secured by mechanisms and processes that replicate all the steps of the check-in process at in-person precinct polling stations, but with the ADDITIONAL safeguard that the voter needs to provide a unique document that was sent to their address for each election.
The idea that mail-in voting invites fraud is simply not supported by reality.
mhoye@owen The part that kills me is that apparently people at Anthropic are reduced to begging the tools to act right just like everyone else, and they're just doing that, illogical nonsense at all. Their whole job is building this thing and they've got no better insight into it or control over it than anyone else.
Alice Averlong🏳️⚧️I remember cynically joking last year that all the "advanced" AIs like Claude were just the same kind of black-box LLMs but with a bunch of regexes glued onto them
now that the source has leaked: HAHAHA
Tim (Wadhwa-)Brown 
Coding with LLMs and agents is a generational opportunity to throw the last decade's hard won lessons on secure coding and appsec out of the window. Definitely something that trust and safety teams, threat actors and possibly even your parents are seizing on with glee when they bypass all of your policies and procedures around installing new software, data governance, validated designs, code reviews, principles of least privilege and regular security assessments. Best of luck.
The future of energy is being decided now. "The US is basically aligning itself as the last big petrostate. We're going to go down with the fossil fuel ship, and China is aligning itself as the first electrostate." https://kottke.org/26/03/us-the-last-big-petrostate
paulrickardsIf only my screen had enough space playing a video in portrait mode for controls to not cover the image. Until that day..
RE: https://mastodon.social/@heidilifeldman/116319275301770100
In addition, remember that these ICE agents are not trained in:
- crowd management
- de-escalation (as required by many states and cities)
- state and local law
- air transport security regulations and procedures
- non-immigration federal law
- investigations
and have had 1/10th the training required by the state with the LOWEST training requirement for certification.
This is intentional. They are there to intimidate, provoke, and take the fall for whatever happens.