Thomas Powers
@thepowers.bsky.social
ML scientist, transit advocate with Seattle Subway
Reposted by Thomas Powers
you know what's better for orcas? If we don't fill the Sound with microplastics. you know what generates microplastics? car tires. you know how you stop people from commuting so much? BY BUILDING DENSER HOUSING
Birds Connect Seattle, Thornton Creek Alliance, and the Orca Conservancy are among a lengthy list of groups trying to send City of Seattle back to the drawing board on its housing growth plan through a judicial appeal in state court.
Story: www.theurbanist.org/2026/01/27/c...
Story: www.theurbanist.org/2026/01/27/c...
Conservation Groups Join Push Against Seattle Growth Plan » The Urbanist
# Birds Connect Seattle, Thornton Creek Alliance, and the Orca Conservancy are among the groups trying to send the City of Seattle back to the drawing board on its housing growth plan. They're pushing...
www.theurbanist.org
January 27, 2026 at 11:06 PM
you know what's better for orcas? If we don't fill the Sound with microplastics. you know what generates microplastics? car tires. you know how you stop people from commuting so much? BY BUILDING DENSER HOUSING
Reposted by Thomas Powers
That feeling when you’re a transit rider and Sound Transit is about to connect Seattle to the Eastside by rail. #thisisyourcity
January 23, 2026 at 4:10 PM
That feeling when you’re a transit rider and Sound Transit is about to connect Seattle to the Eastside by rail. #thisisyourcity
Why don't well-respected people like our plagiarism machines and CSAM generators?
Calling all the PhDs I know to make monumental annoyances of themselves (and honestly, who better)
January 23, 2026 at 2:35 AM
Why don't well-respected people like our plagiarism machines and CSAM generators?
Reposted by Thomas Powers
In our conservative estimate, 41 percent of Grok’s images — or 1.8 million — were sexualized depictions of women. Separately, CCDH analyzed how many were sexualized images across genders and ages. Their findings: an estimated 3 million images, including more than 23,000 of children.
January 22, 2026 at 3:01 PM
In our conservative estimate, 41 percent of Grok’s images — or 1.8 million — were sexualized depictions of women. Separately, CCDH analyzed how many were sexualized images across genders and ages. Their findings: an estimated 3 million images, including more than 23,000 of children.
January 22, 2026 at 2:01 AM
Oh hell yeah
We are beyond honored to have joined Mayor Wilson this morning to announce an executive order for bus lanes on Denny Way. We can't wait to finally Fix the L8 and see what she has in store for the next 4 years of Seattle's transit
January 15, 2026 at 10:39 PM
Oh hell yeah
Reposted by Thomas Powers
“In my mind, the Seattle Transit Measure is about funding transit service. And so I know there’s a tendency to want to put a bunch of other stuff in there. I really do want to make sure that we are prioritizing getting our service levels and reliability back up.”
We sat down with Mayor Katie Wilson to get the latest on initiatives she has cooking now that her team is in office. We touched on emergency housing, bus lanes, World Cup preparations, Sound Transit, culture change at the police department, and more.
Story: www.theurbanist.org/2026/01/13/o...
Story: www.theurbanist.org/2026/01/13/o...
One Week In, Katie Wilson Charts the Path Ahead » The Urbanist
# The Urbanist recently sat down with Seattle Mayor Katie Wilson to get the latest on initiatives she has cooking now that her team is in office. We touched on emergency housing, bus lanes, World Cup ...
www.theurbanist.org
January 15, 2026 at 4:03 PM
“In my mind, the Seattle Transit Measure is about funding transit service. And so I know there’s a tendency to want to put a bunch of other stuff in there. I really do want to make sure that we are prioritizing getting our service levels and reliability back up.”
Reposted by Thomas Powers
Help pass this bill to improve safety in two minutes:
Go to app.leg.wa.gov/csi/House?se...
1) Select the agenda item HB 2095
2) Click that you want your position noted for the record
3) Sign in “pro” and fill out your info
Done! You just helped pass a bill that will keep people safer
Go to app.leg.wa.gov/csi/House?se...
1) Select the agenda item HB 2095
2) Click that you want your position noted for the record
3) Sign in “pro” and fill out your info
Done! You just helped pass a bill that will keep people safer
Hey Washington State people who care about safer streets-- Please voice your support for Washington to lead the US towards safer streets by passing a Presumed Liability Law.
The House Civil Rights & Judiciary Committee is holding a hearing at 8am tomorrow, 1-14-2026 regarding HB 2095.
The House Civil Rights & Judiciary Committee is holding a hearing at 8am tomorrow, 1-14-2026 regarding HB 2095.
January 14, 2026 at 4:02 PM
Help pass this bill to improve safety in two minutes:
Go to app.leg.wa.gov/csi/House?se...
1) Select the agenda item HB 2095
2) Click that you want your position noted for the record
3) Sign in “pro” and fill out your info
Done! You just helped pass a bill that will keep people safer
Go to app.leg.wa.gov/csi/House?se...
1) Select the agenda item HB 2095
2) Click that you want your position noted for the record
3) Sign in “pro” and fill out your info
Done! You just helped pass a bill that will keep people safer
Reposted by Thomas Powers
We don't need a more professional gestapo patrolling our streets.
I got this press release from CM Bob Kettle. What a joke. He’s fine with ICE’s mission, he just thinks they should be more professional about it. We need better leaders than this.
January 13, 2026 at 7:04 PM
We don't need a more professional gestapo patrolling our streets.
Reposted by Thomas Powers
This is simply not technically possible. You cannot generate enough writing in-house to train an LLM. You cannot generate enough art in-house to train a GAN. People who think they have a model "trained on our own material" are using a base model of mass-scraped data, unknowingly or disingenuously.
Larian CEO says Divinity could have gen AI assets in it but only if the tools are trained on the studio's own data.
"If we use a GenAI model to create in-game assets, then it'll be trained on data we own."
https://kotaku.com/divinity-gen-ai-larian-bg3-reddit-ama-2000658429
"If we use a GenAI model to create in-game assets, then it'll be trained on data we own."
https://kotaku.com/divinity-gen-ai-larian-bg3-reddit-ama-2000658429
January 9, 2026 at 3:39 PM
This is simply not technically possible. You cannot generate enough writing in-house to train an LLM. You cannot generate enough art in-house to train a GAN. People who think they have a model "trained on our own material" are using a base model of mass-scraped data, unknowingly or disingenuously.
January 9, 2026 at 1:29 AM
Reposted by Thomas Powers
Abolish and prosecute ICE. They did a murder in broad daylight today. Imagine what they're doing in the closed facilities we're somehow not even allowed to see.
January 7, 2026 at 6:20 PM
Abolish and prosecute ICE. They did a murder in broad daylight today. Imagine what they're doing in the closed facilities we're somehow not even allowed to see.
Reposted by Thomas Powers
I realize this is a story from like 50 years ago in internet time, but I wrote about how it was journalistic malpractice for major news orgs to claim that "Grok apologized" when it literally CANNOT do that, and there is no sign that xAI has made a single change to its deepfake porn generator.
Journalistic Malpractice: No LLM Ever ‘Admits’ To Anything, And Reporting Otherwise Is A Lie
Over the past week, Reuters, Newsweek, the Daily Beast, CNBC, and a parade of other outlets published headlines claiming that Grok—Elon Musk's LLM chatbot (the one that once referred to itself as…
Over the past week, Reuters, Newsweek, the Daily Beast, CNBC, and a parade of other outlets published headlines claiming that Grok—Elon Musk's LLM chatbot (the one that once referred to itself as…
Journalistic Malpractice: No LLM Ever ‘Admits’ To Anything, And Reporting Otherwise Is A Lie
Over the past week, Reuters, Newsweek, the Daily Beast, CNBC, and a parade of other outlets published headlines claiming that Grok—Elon Musk's LLM chatbot (the one that once referred to itself as "MechaHitler")—had "apologized" for generating non-consensual intimate images of minors and was “fixing” its failed guardrails. Grok did no such thing. Grok cannot apologize. Grok is not a human. Grok has no sense of what is happening.
www.techdirt.com
January 7, 2026 at 7:38 PM
I realize this is a story from like 50 years ago in internet time, but I wrote about how it was journalistic malpractice for major news orgs to claim that "Grok apologized" when it literally CANNOT do that, and there is no sign that xAI has made a single change to its deepfake porn generator.
Are there any editors at these legacy MSM outlets who have basic tech literacy? This should be disqualifying.
This is a thread of major media outlets falsely anthropomorphising the "Grok" chatbot program and in doing so, actively and directly removing responsibility and accountability from individual people working at X who created a child pornography generator (Elon Musk, Nikita Bier etc)
#1: Reuters
#1: Reuters
January 3, 2026 at 1:02 AM
Are there any editors at these legacy MSM outlets who have basic tech literacy? This should be disqualifying.
Reposted by Thomas Powers
1. Headlines everywhere today read "Grok apologizes."
This is bullshit. A chatbot is not something that can apologize.
Pretending otherwise is simple laundering these companies' bullshit about what AI is, while diffusing blame away from the human beings that developed and released this system.
This is bullshit. A chatbot is not something that can apologize.
Pretending otherwise is simple laundering these companies' bullshit about what AI is, while diffusing blame away from the human beings that developed and released this system.
January 3, 2026 at 12:12 AM
1. Headlines everywhere today read "Grok apologizes."
This is bullshit. A chatbot is not something that can apologize.
Pretending otherwise is simple laundering these companies' bullshit about what AI is, while diffusing blame away from the human beings that developed and released this system.
This is bullshit. A chatbot is not something that can apologize.
Pretending otherwise is simple laundering these companies' bullshit about what AI is, while diffusing blame away from the human beings that developed and released this system.
December 31, 2025 at 4:30 AM
Reposted by Thomas Powers
In my case, this painstaking labor in the process through which the science gets done. Science isn't measuring stuff in a lab. It's thinking deeply, extracting the heart of idea from the soup of thoughts running through my mind, molding it, and finding a way to communicate that idea to others.
December 16, 2025 at 11:46 PM
In my case, this painstaking labor in the process through which the science gets done. Science isn't measuring stuff in a lab. It's thinking deeply, extracting the heart of idea from the soup of thoughts running through my mind, molding it, and finding a way to communicate that idea to others.
Reposted by Thomas Powers
My least popular (and most correct) view is that cars should be automatically limited to the local speed limit. Put the pedal to the floor and you still can't go over 25mph in a residential area.
(15 in Manhattan btw)
(15 in Manhattan btw)
There is no possible justification for limiting e-bikes to 15mph but not cars.
New York Has a New E-Bike Speed Limit—and No Way to Enforce It
August 9, 2025 at 4:47 PM
My least popular (and most correct) view is that cars should be automatically limited to the local speed limit. Put the pedal to the floor and you still can't go over 25mph in a residential area.
(15 in Manhattan btw)
(15 in Manhattan btw)
Reposted by Thomas Powers
…in case you’re wondering what’s going on behind the scenes with shared streets…
A Sneak Peek at Shared Streets Coming to Seattle » The Urbanist
# A trove of documents obtained from City of Seattle through public disclosure requests shows work is proceeding gradually but steadily inside the transportation department to take advantage of Washin...
www.theurbanist.org
December 4, 2025 at 9:36 PM
…in case you’re wondering what’s going on behind the scenes with shared streets…
November 29, 2025 at 12:53 AM
Even ignoring the obvious AI slop figure, the paper is essentially applying an existing model to 1 (one) tiny open dataset. Taking it at face value, it reads like an undergrad take-home assignment that would get an F
"Runctitiononal features"? "Medical fymblal"? "1 Tol Line storee"? This gets worse the longer you look at it. But it's got to be good, because it was published in Nature Scientific Reports last week: www.nature.com/articles/s41... h/t @asa.tsbalans.se
November 28, 2025 at 4:49 PM
Even ignoring the obvious AI slop figure, the paper is essentially applying an existing model to 1 (one) tiny open dataset. Taking it at face value, it reads like an undergrad take-home assignment that would get an F
Reposted by Thomas Powers
Wow this is bad.
CPSC Warns Consumers to Immediately Stop Using Batteries for E-Bikes from Rad Power Bikes Due to Fire Hazard; Risk of Serious Injury or Death www.cpsc.gov/Warnings/202...
November 24, 2025 at 7:16 PM
Wow this is bad.
Now's the time of year I have to bike home in the dark, but at least I get to see cool sunsets over the water
#photography #pnw #seattle
#photography #pnw #seattle
November 14, 2025 at 3:25 AM
Now's the time of year I have to bike home in the dark, but at least I get to see cool sunsets over the water
#photography #pnw #seattle
#photography #pnw #seattle
Reposted by Thomas Powers
redzone showing that
November 9, 2025 at 11:25 PM
redzone showing that