Why am *I* the weird one for distrusting the computer that catfishes you?
Remote
dreid
@dreid@wandering.shop
mastodon 4.7.3Rollerblader (Wizard Skater), hand tool woodworker, and software engineer in Alameda, CA.
550 Followers
367 Following
42 Posts
Joined November 22, 2022
Pronouns:
He/him
Useful Website:
Empty Website:
Signal:
dreid.01
@1password hey what's up with sponsoring DHH's vanity "Linux distro"?
Are you somehow not aware of his history of racist blog posts?
In case that's the case here you go:
An overview: https://jakelazaroff.com/words/dhh-is-way-worse-than-i-thought/
Lest you think those quotes were somehow out of context: https://world.hey.com/dhh/wolves-sheep-and-gypsies-ba44af6a
https://world.hey.com/dhh/three-sacred-cows-that-must-die-so-europe-can-live-1afb203d
https://world.hey.com/dhh/as-i-remember-london-e7d38e64
Edit: if you would like to voice your displeasure with 1Password more directly you should email support@1password.com
It's really not that hard to live your life such that when you die millions of people *DON'T* cheer.
It's basically the default.
A developer ecosystem optimized for LLM usage is as bad as a city optimized for cars.
Therac-25 huh? That shit is wild. I can't believe industry would be so irresponsible.
Anyway time to ask my Markov Waifu to make a list of probable military installations to bomb.
Replying to
@Corax42@mastodon.social @_elena@mastodon.social in this case one does not even have to assume.
Jason Meller the VP of Product of at 1Password is a big fan of DHH.
For years https://how.complexsystems.fail has been missing a link for one of the documents in the "Other materials" section.
Specifically "Woods & Cook (1998). Characteristics of Patient Safety: Five Principles that Underlie Productive Work. (Chicago: CtL)"
I would search for it periodically but was never able to find it.
But this weekend I met jek at #nbpy who after learning about this one missing citation was able to track it down and send it to me.
All the links are finally blue.
There are no "good guy" AI companies.
You can't use any of this shit without enriching the worst fucking people.
Fucking wild to me that Pro-LLM people will pretend not to understand the difference between a person torrenting Dr Who to watch and the richest corporations on the planet hoovering up all information to feed into plagiarism machines.
There is something deeply offputting about watching engineers who have made cool stuff in the before times desperately pleading with models not to make everything worse.
Canonical's ability to do something stupid that will annoy their users and also being late to the hype party is remarkable.
Several years ago when we first started using typescript we stuck [key: string]: any onto the Vue interface and it has been hiding hundreds of type errors ever since. Now finally in the last mile of what is hopefully the final attempt to migrate from Vue2 -> Vue3 and fuck it I'm just gonna have to fix all of them.
(Anyone who suggests using an LLM even in jest is getting immediately blocked.)
Pretty fucked up how people are apparently more willing to attribute autonomy and agency to their rented markov bot girlfriends than to trans people.
Replying to
@paco@infosec.exchange @pluralistic@mamot.fr similarly no one has announced "thanks to AI we have moved to a 4 day workweek."
Part of the point of open source is reuse. And wholesale copying can be valid. But that reuse has always had boundaries and constraints applied to it.
Attribution, open sourcing of modifications, etc.
Thats what the licenses are there for.
If we're saying licenses shouldn't matter then ok that's an argument we can have.
But the "clean room implementation of a thing in the training data" people still seem to think licenses matter or they wouldn't have had to relicense anything.
Python 4 but all it does is make `foo = "bar",` a syntax error instead of a tuple.
"DREAM stands for Driving Ruby's Evolution to AI Maturity."
lol and I can not stress this enough lmfao
Replying to
Anthropic's public fight with the DoD about being unwilling to use their technology in hypothetical murder bots is probably less a hard ethical line and more a savvy PR move to distract from all the other harms to society the technology is facilitating.
Every app is an offline first app if your vendors are unreliable enough.
Today is my birthday. My wife got me the complete hitchhiker's guide trilogy.
I bet you can't guess how old I am.
If you'd like to get me something I am proud to share my birthday with trans day of visibility. So perhaps make a donation to https://tcpipeline.org/
https://code.visualstudio.com/docs/debugtest/integrated-browser
The editor in a web browser now has a web browser in the editor.
I don't care about rockets
I don't care about moon
I hope the thing we're all waiting for
Happens really soon
As a southerner I feel like I should plug Dixie Be Damned for some stories of resistance that don't end up in most history books.
Replying to
@mhoye@cosocial.ca yeah it's been in my bio for a while but I don't really talk about it much. (It doesn't require much work to maintain). I'm glad people find it useful.
I wonder how much software developer culture is segmenting based on what kind of futzing people enjoy and maybe we're learning that more people than expected would rather be writing emails than code.
Replying to
Yes they were just complying with a legal warrant and that is the cost of doing business in a jurisdiction.
If they hadn't complied they'd probably just get popped by the NSA.
They've still spent decades overselling the privacy capabilities of GPG+Switzerland.
Also you should probably assume they've been popped by the NSA.
Ronan Farrow's New Yorker article about how Sam Altman can't be trusted has a slop animation above the fold.
I am so fucking tired.
Replying to
@ceejbot @glyph @aredridel there is the assumption that they won't make bad code forever so we'll just use them to replace the bad code with better code later.
Also we can put the person who committed the bad code on a PIP while ignoring the organizational defects that prevented putting any safety procedures into place.
On paper there is lots of interesting stuff happening. But in practice I can't figure out a way to use it without enriching the worst fucking people.
Replying to
@glyph @aredridel @ceejbot I think a lot of people are working on the assumption that mistakes aren't as costly anymore.
You won't have to live with the consequences very long and you can just rewrite everything if the technical decisions you make end up being wrong.
This doesn't hold for genuine safety issues, like things affecting the privacy and security of your users, but industry was already caring about those things pretty reluctantly.
Replying to
@jacob@social.jacobian.org @glyph@mastodon.social I prefer to think of it as aspirational.
Replying to
@glyph @aredridel @ceejbot given the choice of being out competed by someone using AI and losing all your customers data because you used AI the choice if obvious.
There are provably no consequences for the latter.
You can calculate exactly how much a year of free credit monitoring for all your users will cost.
Replying to
@hynek our orm avoidant codebase somehow replicated all the problems (n+1 queries everywhere, etc) with none of the "benefits".
We do have explicit transaction management with separate DBReadTransaction and DBWriteTransaction types though it's nice.
Replying to
@eri Camp Indigo a summer day camp for trans and non-binary youth.
https://www.thecampindigo.org/
Replying to
After appeals to the governor. Probably on the burner phones he sent them. https://www.politico.com/news/2025/03/18/newsom-ceos-burner-phones-00235044
Replying to
@mik@infosec.exchange It was only a little bit of a joke though.
I've definitely had my "this is easier if I just write OpenSCAD" moments.
Not G-code so far though.
Replying to
@hynek it's pretty bad. Could be worse. I've seen non-SQL database clients that implement __del__ to perform remote deletes.
There are CNC people out there just raw dogging G-Code.
Now that's what I call close to the metal.