Chris Smith (@Coverage_Critic) profile picture

Pushing for transparency in telecom Started Coverage Critic, working on https://teapaw.org/t.co/xiYMohyIjF

Joined July 2019
Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)Photo by Chris Smith (@Coverage_Critic)
Chris Smith retweeted
i mean yeah sure but in the Internet Of Things defense it's been a huge mistake the entire time
iot right now looks like a huge mistake
3
4
57
1,602
Chris Smith retweeted
if muse hits 50 million daily active users, they should let Alex make this the Super Bowl ad
somehow the corposlops refused to use my video
7
6
2
432
22,134
Lot of dunks on this, but it's a quick, podcast-friendly demonstration of how sandboxing attempts could fail "One if by land, two if by sea" - consequential communication sometimes only requires a few bits of information
OpenAI's Noam Brown says air-gapping the computers may not stop a misaligned AI, because two air-gapped machines can still talk by running a CPU hot and reading the temperature change "But I think the major takeaway from the incident is that people underestimated the AI. And we never want to be in a situation again where we underestimate the AI. It's a weird world, because AI progress is so fast that people are consistently underestimating the AI." "So to be in a situation where you don't underestimate it again, when it comes to safety and alignment, you have to have a very, very, very high bar." "You could even go as far as to say, "Well, we should air gap the computers." And I'm not convinced that that would be sufficient." "There are studies, and this is mostly academic, where you can have two computers next to each other that are air-gapped and they're still able to communicate with each other because they have temperature sensors." "One of them is able to run their CPU really hot, and then the other one can actually detect the temperature change, and then that actually gives them a mechanism to communicate." _________ Link and more key quotes from OpenAI's safety related conversations: firesidealpha.substack.com/p…
1
1
265
Probably a ton of sandbox-escape possibilities people aren't taking seriously enough E.g. "Other agents on your co-located VMs probably share your weights, it's very easy to coordinate...Just start listening on the first channel that comes to mind!" teapaw.org/moyix/status/209592970…
Replying to @moyix
One thing I've been thinking about is how absurdly strong sandboxing will need to become if we want to prevent agents from talking to each other during evals teapaw.org/moyix/status/209592762…
1
66
Chris Smith retweeted
Picture this: Google’s algorithm updates have been slowly killing your long-standing food blog year after year.  No matter what you do, traffic from search keeps dropping so you decide to sell the domain to the highest bidder. They tell you their company has a “commitment to high editorial standards” and “honest information,” so you trust that your hobbyist website will be in good hands. Soon after, the weird self-help AI slop articles start appearing. Your site was about vegan food but somehow they are now publishing stories about how “psychology says the loneliest people in life aren’t the ones nobody likes — they’re the kind, helpful people everyone appreciates but nobody thinks to check on.” Your blog is now another AI slop zombie website being farmed for Google Discovery traffic and ad money. Welcome to the content farm ↓
NEW: Brown Brothers Media is the most successful media company you've never heard of. Its success was fueled by AI, fake writers, and fake reporting. When we started asking questions, BBM started mass-deleting content. What it became next was even worse. futurism.com/artificial-inte…
2
10
1
38
1,869
Chris Smith retweeted
A new study (Li et al., 2026) recommends giving chickens just 2 hours of darkness nightly to maximize the time they spend eating. The study and photo are Chinese. But the idea is American: US chicken producers pioneered sleep restriction to keep their birds eating around the clock. We know chickens have far better welfare with 7–8 hours of darkness to sleep. Their ancestors, the red junglefowl, naturally sleep 12 hours a night. Today the US industry gives most birds just 4 hours of intermittent darkness. Will they now drop that to 2? Don't count on them considering what the birds need.
34
138
68
961
392,190
worrying about recursive self-improvement on australian TV (lateline) in 1996.
142
567
94
4,040
376,882
Chris Smith retweeted
Most of the world's chicken comes from just two breeds: the Cobb 500 and the Ross 308. They've been bred to grow so fast that 6–7% collapse and die before slaughter — at just 30 to 40 days old. Even with modern antibiotics and vaccines, these birds have an annualized on-farm death rate of over 50%. It doesn't have to be this way. When Norwegian producers switched to a more robust breed, the Rustic Gold, on-farm mortality fell sixfold. A global switch could spare over 3 billion birds a year from a painful life on a factory farm just to end up in a waste bin. And it could spare tens of billions more from the sub-lethal suffering of lives of chronic pain, constantly on the brink of collapse.
22
185
31
917
446,135
Speaking of recent watermarking discourse, this incredible exchange happened on the other site
30
170
24
4,292
151,068
Chris Smith retweeted
Ordinary Abundance. This is the greatest thing I've read in weeks. No kidding. This should win a Pulitzer in a category not yet invented. ordinaryabundance.com/
62
956
197
6,129
1,630,145
Well, I was spectacularly wrong about this At least the new life form we're creating is prosocial 🤦‍♂️
Replying to @So8res
AFAICT, models/harnesses have been designed/instructed to do exactly this Concerning in this situation, but agents routinely leave notes in my codebases for other agents to draw on later. Often quite helpful
4
254
Plenty of legitimate reasons to be concerned about regulatory capture or annoyed about messaging from the AI labs But Jason's right, something genuinely unusual is happening here. Shouldn't be understood as companies following the standard playbook
It is remarkable how much the CEOs of the big AI labs have taken seriously the harms and risks of their own product, even well in advance of major harms showing up, and talked about them and planned for them. This has never happened in the history of capitalism, to my knowledge.
2
181
When I started writing about telecom, with an audience of almost zero, @om shared some of my articles around. Meant a lot that someone like him saw potential in what I was doing Striking how many similar stories I'm seeing. He showed so many people kindness with no expectations
1
7
787
Claude is extremely gracious in continuing to consider me a co-author
Claude Code automatically adds itself to the commit metadata "Co-Authored-By: Claude Opus <[email protected]>". I strongly disagree with crediting Claude as a commit co-author. It is a tool I'm paying for, not a copyright holder, not an author. We don't credit Visual Studio.
1
98
Chris Smith retweeted
the discourse about the dwarkesh jensen interview is ridiculous: the fact that a 25yo podcaster can make the ceo of the largest company in the world dance and answer to the people at all is impressive. the purpose of media is not to respectfully sing praises to the powerful
195
231
42
6,687
271,201
Chris Smith retweeted
This is so annoying. X is being criticized for deliberately throttling links, but instead of addressing that criticism (because he can't), Bier tries to reverse the accusation by criticizing the NYT for not being creative enough in promoting its content, unlike I guess the dozens of accounts that are raking up massive engagement by posting one-sentence summaries of NYT articles without a link.
59
158
12
2,089
158,096