A saxophone someplace far-off played

As she was walkin' on by the arcade

As the light bust through a beat-up shade Where he was waking up She dropped a coin into the cup Of a blind man at the gate And forgot about a simple twist of fate

Gorgeous: rural Kyoto in the heavy snow.

Am I a sucker for these kinds of ultra-detailed images by Richard Nadler because I am a fan of Richard Scarry books and Wes Anderson movies or am I a fan of Scarry & Anderson because I’m a sucker for these kinds of ultra-detailed images? (Or is it because I’m aphantasic and require external imagery for this level of detail?)

See also Mark Alan Stamaty’s NYC Illustrations and Infinite Illustrations.

KDO: 28 Years Later…

It’s getting a little ridiculous, isn’t it? 28 years of kottke.org, as of today. Older than Google. Older than The Matrix. Older than Christopher Nolan’s feature film career. Older than Elle Fanning. Older than Kurt Cobain when he died. 47,300 posts since March 14, 1998. It might outlast American democracy.

KDO retains its old school vibe but with some new tricks. Friends and readers have remarked recently that I seem to be having fun with the site again and that’s true. Still excited by the possibilities of what the site could become. I hope you’ve been enjoying it.

I’d like to once again thank KDO members for supporting the site — I never get tired of the member thanking. They each pay a few dollars a month to keep the site free to read for everyone, regardless of income, something I feel very strongly about in this era of paywalled media. As Hamilton Nolanhas noted — his site operates on the same principle — sites/newsletters like these have progressive funding structures, where people’s contributions are based on their ability to pay to help keep the site available to all. If you’d like to help support the site in this mission, you can check out your membership options here. ✌️

P.S. I hope you enjoy the birthday logo and fireworks — they’ll be around until Monday.

DOJ clears way for government to hire technologists still connected to private sector employers

NextGov/FCW – “The Justice Department issued an opinion last week authorizing the Trump administration’s plan to allow employees from tech companies to work for the federal government while remaining employed by their companies and keeping their not-yet-vested company stocks. The administration will be onboarding managers from twenty-plus companies including Anduril, Microsoft, Nvidia, OpenAI, Palantir and xAI as part of its U.S. Tech Force program, launchedlast year to recruit early-career engineers after the administration pushed over 20,000technologists out of their government posts last year.  

The setup is an unusual one. Federal employees are subject to ethics rules meant to ensure that they work for the public interest. The Office of Personnel Management now has DOJ’s blessing to allow individuals joining the Tech Force to keep their restricted stock units that haven’t yet vested company stocks issued with a vesting plan that dictates when employees get full ownership of them while they work for the government on a leave of absence from their private sector employer.  Ethics experts and public sector lawyers told Nextgov/FCW that they are skeptical about the arrangement. 

 “Why are we replacing a workforce we already had with individuals who may still be beholden to an outside employer?” asked Cynthia Brown, the senior ethics counsel at the nonprofit watchdog organization Citizens for Responsibility and Ethics in Washington. It “raises a lot of very serious concerns.” It’s unclear exactly how many people will join the government on a leave of absence as part of the Tech Force. It’s likely that the DOJ decision will be primarily used for the 100-plus managers being recruited from tech companies partnering with the government, rather than for the class of early-career employees, an OPM spokesperson told Nextgov/FCW. 

 “This opinion from the Department of Justice provides much-needed clarity on the treatment of deferred compensation and strengthens the federal government’s ability to recruit top talent from the private sector to complete stints of government service,” the spokesperson said in a statement. The statute addressed by DOJ in its opinion generally bans federal employees from receiving outside compensation for their government service…”

• CGP has a record to all the OLC memos/slip opinions back to 1934 when it began producing them (OCLC) 839756075, System Number 000599089. memo [h/t Ben Amata]

The Anthropic Institute

“From inside a frontier AI lab, we confront the most significant challenges about how powerful AI will impact the world around us.. The Anthropic Institute exists to understand and shape the consequences of powerful AI systems. 

We focus on the urgent questions that will determine whether these systems deliver the radical upsides that we believe are possible in science, security, economic development, and human agency—or whether they will pose a range of unprecedented new risks to humanity. The Institute works across four major challenges, conducting technical work to understand how AI systems behave with respect to each, and how our societies will need to respond.” One example of this research is the following report: Anthropic Economic Index Understanding AI’s effects on the economy. Last updated: Jan 15, 2026.

And Related – See also via Slashdot: “In 2024 Anthropic was sued over claims it infringed copyrights when training LLMs. But as they try to settle, they may have a problem. The Free Software Foundation announced Friday that Anthropic’s training data apparently even included the book “Free as in Freedom: Richard Stallman’s Crusade for Free Software” — for which the Free Software Foundation holds a copyright. It was published by O’Reilly and by the FSF under the GNU Free Documentation License (GNU FDL). This is a free license allowing use of the work for any purpose without payment.

 Obviously, the right thing to do is protect computing freedom: share complete training inputs with every user of the LLM, together with the complete model, training configuration settings, and the accompanying software source code. Therefore, we urge Anthropic and other LLM developers that train models using huge datasets downloaded from the Internet to provide these LLMs to their users in freedom. We are a small organization with limited resources and we have to pick our battles, but if the FSF were to participate in a lawsuit such as Bartz v. Anthropic and find our copyright and license violated, we would certainly request user freedom as compensation. “The FSF doesn’t usually sue for copyright infringement,” reads the headline on the FSF’s announcement, “but when we do, we settle for freedom.”