Pinned
Chris Riley
Chris Riley
@mchrisriley.com
"A blend of tech insights, dad humor, and existential musings... Bonus points if you can explain data portability over bourbon." http://mchrisriley.com (... of course)
2.2K followers1K following491 posts
New from me at Tech Policy Press: Building Trust Infrastructure for Agentic AI. There are challenges coming; let's get out ahead of them, if it's not already too late. www.techpolicy.press/building-tru...
Chris Riley reposted
Fantastic thread linking Taiwanese independence, the state of the chip market, and the stability of the US economy.
AI drives enormous demand for high-performance chips; I won't get into the technical details here, unless people want. What's important to keep in mind is that, of the actual chip market, there are three companies that can make them: TSMC, Samsung, and Intel. None of those are Chinese.
Excited for this event coming up on Friday: "Who Controls Data in the Age of AI?" luma.com/kpf2voda Sign up and come to UC Law San Francisco to learn about the law and practice of data transfers with several distinguished speakers, and me. (But I organized it, so they had to include me, mwahaha!)
We are, collectively, struggling to come to terms with the emerging capabilities of AI. Two conversations are happening with no cross-pollination, optimistic and pessimistic. I am trying to create a workable theory to reach both. Here is my chapter on self: mchrisriley.medium.com/chapter-9-th...
Experiments conducted with the A.I. system Claude are producing fascinating results—and raising questions about the nature of selfhood. Gideon Lewis-Kraus reports from inside the company that designed it, Anthropic. newyorkermag.visitlink.me/rOfXjg
"Doing harm at scale requires infrastructure." So good, from @johnathan.bsky.social @shappy.bsky.social I always read World's Best Newsletter, but this one I am going to save and reread. mailchi.mp/rawsignal/yo...
Chris Riley reposted
Federal judge has ordered the release of Liam Ramos (the bunny hat boy) and his father. A brief and rather remarkable order. Clearly written to be shared widely, so please do. storage.courtlistener.com/recap/gov.us...
Chris Riley reposted
NEW: When can states prosecute federal agents? While it may not be easy, especially if federal officials impede an investigation, states have more power than people may realize. @aliciabannon.bsky.social breaks down the tools states have to hold federal officials accountable:
Chris Riley reposted
Last week in Minnesota, I watched ordinary people risk their lives to protect their neighbors. In the process, they not only won a significant—though not final—victory against authoritarianism, they proved virtually every MAGA social theory wrong. (gift link) www.theatlantic.com/ideas/2026/0...
www.nytimes.com/2026/01/23/s... There is a callout box in the print version of this piece that reads, "Like all gods, this technology is inscrutable." .... sheesh. The piece itself is pretty great though, and reinforces what I wrote about idolatry and dangerous mutual worship dynamics.
Excited to kick off the spring concert prep cycle!
Our 25th Anniversary Season has been busy so far, and we aren’t done! In just a few short weeks, we begin rehearsing for our May concert! Stayed tuned for more information about this concert because it’ll feature some amazing compositions and musicians!
This is an incredible conversation, touching on several distinct and worthy issues -- not just AI, but even more so science and how we approach it, the pervasive confusion of correlation and causation, and a bit of current politics/events at the end. Well worth a read.
"Maybe there will be artificial sentience. But I don’t think any sane person thinks large-language models are that. They don’t have inner autonomous agency. They’re a tool that you either use or you don’t. They're not intelligent machines that have lived experiences in the world." - Jessica Riskin
I'm sure literally no one involved in making this happen will care, but I shudder to think of the child psychologist reaction to this Universal Basic Investment proposal: www.axios.com/2025/12/21/t... Making every kid aware of, and competing over, money seems like a fast way to ruin childhood.
As a practicing Christian, I genuinely believe God engages with the present world in mysterious ways, and if we but listen quietly and humbly, He will speak to us. ... but not like that. Gift link: www.nytimes.com/2025/11/26/o... (See also, by me: mchrisriley.medium.com/interlude-co...)
Chris Riley reposted
Chris Riley reposted
This is, of course, a fascinating thread from Kate asking why Google is effectively cannibalizing its own market for both digital ads on third party properties and search ads, by effectively taking away the incentive to click through to third party sites. I have a theory as to why they're doing it.
So this ad revenue metric is indicative of such a fascinating and weird set of business decisions by Google to inexplicably kill of SEO-- and I honestly don't get how they think this is going to pay off for them in the long run. 🧵 1/
Today is @dansinker.com 's birthday. If you're not already signed up for his emails, you should! dansinker.com/follow/ And maybe buy some stickers or something. I love my "TRYING" sticker, it is the anchor sticker on my laptop, the first one I put on.
Oh and before you ask, "Wow, Chris, what are you writing??" Here is the introduction: mchrisriley.medium.com/introduction... - Chapter 1 is up too, and I'm in the midst of editing Chapter 2. It's extra ironic to use an AI copyeditor to write about how AI is absolutely not conscious, and not close.
AI has no place in beer. Today's edition of "sentences I never thought I'd have to say", but thanks to @404media.co apparently we do? www.404media.co/what-happene...
The longevity cult is going to be ALL OVER the naked mole rats. For science, I guess. Hat tip @404media.co of course, because they are the best.
I really enjoyed this piece on AI + consciousness, a topic I've been thinking about a lot lately. It covers some of the key thinkers + dimensions of the problem, and touches on why it matters: the future possibility of moral welfare, + the risk of it distracting from more pressing concerns.
"The human brain is simply not wired to treat AI like any other technology," Webb Wright writes. "For some users, the system is alive."
Chris Riley reposted
cool cool apnews.com/article/unga...
Chris Riley reposted
Governing AI starts with user control To keep AI open, competitive, & user-centred, we need real data portability — no more platform lock-in, writes @mchrisriley.com, Data Transfer Initiative, in a new blog: www.bennettschool.cam.ac.uk/blog/governi...
👀 And don't miss the closing line of their announcement: "If you would like to bring your memory details over from a different AI tool or export your memory from Claude for backup or migration, you can follow these instructions." (link goes to: support.anthropic.com/en/articles/... )
Anthropic’s Claude AI can now automatically ‘remember’ past chats
... wow
Today, Sen. Ted Cruz proposed the "Strengthening Artificial intelligence Normalization and Diffusion By Oversight and eXperimentation Act," or SANDBOX Act. The bill would create a mechanism for AI developers to seek to waive or modify federal regulatory requirements for a period of up to ten years.
Chris Riley reposted
Chris Riley offers a set of principles for personal AI data transfers, and proposes that technologists begin to immediately develop AI data transfer tools, starting with a concrete and yet valuable data model: generative AI conversation histories.
Another excellent piece from Rosie Spinks: rojospinks.substack.com/p/does-where... Read through to the end, the bit about how AI will remove more friction, which will undercut our thriving and sense of life's meaning. Great stuff.
And yet, there will be (many) mainstream Democrats hemming and hawing about whether this is an intellectually sound or fair argument to make, on the grounds that allegedly Trump and friends are paying for the costs, as if that distinction mattered any more.
“The president cut your health care so he could build himself a golden ballroom” is a pretty easy message to sell
I can see this, particularly in 2025, yeah. But my greater concern is that we will put AI in charge of everything, and then everything will break.
As I told a friend of mine recently, I'm not concerned with AI becoming intelligent (let alone "super intelligent"), I'm concerned humans will make laws saying I have to do what AI tells me to do (or almost worse, AI used in ways that can endanger my life without me knowing it).
Sam Altman says "don't use GPT for therapy" (techcrunch.com/2025/07/25/s...), meanwhile "therapy/companionship" is literally the #1 use case (hbr.org/2025/04/how-...)