Peter Swimm
@peterswimm.com
Founder @toilville
https://www.toilville.com/
https://www.toilville.com/
Reposted by Peter Swimm
November 25, 2025 at 10:14 PM
Reposted by Peter Swimm
Resisting GenAI and Big Tech in Higher Education - Tuesday 25th Nov, free registration, great panel, big focus on climate justice. Register here: lmula.zoom.us/webinar/regi...
November 21, 2025 at 2:02 PM
Resisting GenAI and Big Tech in Higher Education - Tuesday 25th Nov, free registration, great panel, big focus on climate justice. Register here: lmula.zoom.us/webinar/regi...
Reposted by Peter Swimm
Okay, I hestitate to even share this link, because while I like NYMag I do not want to send this journalist in particular clicks, b ut:
If your framing is that an academic is the "dominant voice" and the underdogs are OpenAI, Anthropic and Google, maybe a fact check in is order??
>>
If your framing is that an academic is the "dominant voice" and the underdogs are OpenAI, Anthropic and Google, maybe a fact check in is order??
>>
Is ChatGPT Conscious?
Many users feel they’re talking to a real person. Scientists say it’s time to consider whether they’re onto something.
nymag.com
November 25, 2025 at 1:40 PM
Okay, I hestitate to even share this link, because while I like NYMag I do not want to send this journalist in particular clicks, b ut:
If your framing is that an academic is the "dominant voice" and the underdogs are OpenAI, Anthropic and Google, maybe a fact check in is order??
>>
If your framing is that an academic is the "dominant voice" and the underdogs are OpenAI, Anthropic and Google, maybe a fact check in is order??
>>
Reposted by Peter Swimm
This is *amazing*. This goes straight to Deceptive Design Hall of Shame.
They made a "take a break" nudge that has no obvious "ok, I'll take a break" affordance. Its three affordances are:
1) Keep chatting (default, highlighted)
2) x out — keeps chatting
3) "This was helpful" — what is this?
🧵
They made a "take a break" nudge that has no obvious "ok, I'll take a break" affordance. Its three affordances are:
1) Keep chatting (default, highlighted)
2) x out — keeps chatting
3) "This was helpful" — what is this?
🧵
One of the changes that OpenAI has made to make ChatGPT safer is a "take a break" nudge. There's something quite interesting about the design here. Which thing does it make you want to click?
November 24, 2025 at 2:17 PM
This is *amazing*. This goes straight to Deceptive Design Hall of Shame.
They made a "take a break" nudge that has no obvious "ok, I'll take a break" affordance. Its three affordances are:
1) Keep chatting (default, highlighted)
2) x out — keeps chatting
3) "This was helpful" — what is this?
🧵
They made a "take a break" nudge that has no obvious "ok, I'll take a break" affordance. Its three affordances are:
1) Keep chatting (default, highlighted)
2) x out — keeps chatting
3) "This was helpful" — what is this?
🧵
Reposted by Peter Swimm
Yeah was also confused by all the glowing reviews. But this one is good. blackgirlwatching.substack.com/p/one-battle...
One Fetish After Another: PTA Exploits Black Women and Averts Revolution
Paul Thomas Anderson's action flick leaves much to be desired.
blackgirlwatching.substack.com
November 24, 2025 at 4:48 PM
Yeah was also confused by all the glowing reviews. But this one is good. blackgirlwatching.substack.com/p/one-battle...
Reposted by Peter Swimm
They can’t make computers as intelligent as humans, but they can make humans as unthinking as computers.
November 24, 2025 at 5:18 PM
They can’t make computers as intelligent as humans, but they can make humans as unthinking as computers.
i’ve never walked out of a movie i was screening in my own home but one battle after another made me consider it.
November 24, 2025 at 4:31 PM
i’ve never walked out of a movie i was screening in my own home but one battle after another made me consider it.
Reposted by Peter Swimm
i'm going to sound sarcastic here but genuinely one of the most interesting aspects of LLMs is that they are proving that evil is real, by which i mean if you train them to do one bad thing they start doing other bad things that you have not trained them to do
At least once a month, Anthropic puts out an alignment paper d AI behavior where, if you saw it in a science fiction film, you‘d be screaming at the idiot scientists onscreen to stop development
www.anthropic.com/research/eme...
www.anthropic.com/research/eme...
From shortcuts to sabotage: natural emergent misalignment from reward hacking
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
www.anthropic.com
November 24, 2025 at 6:23 AM
i'm going to sound sarcastic here but genuinely one of the most interesting aspects of LLMs is that they are proving that evil is real, by which i mean if you train them to do one bad thing they start doing other bad things that you have not trained them to do
Reposted by Peter Swimm
At least once a month, Anthropic puts out an alignment paper d AI behavior where, if you saw it in a science fiction film, you‘d be screaming at the idiot scientists onscreen to stop development
www.anthropic.com/research/eme...
www.anthropic.com/research/eme...
From shortcuts to sabotage: natural emergent misalignment from reward hacking
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
www.anthropic.com
November 24, 2025 at 5:54 AM
At least once a month, Anthropic puts out an alignment paper d AI behavior where, if you saw it in a science fiction film, you‘d be screaming at the idiot scientists onscreen to stop development
www.anthropic.com/research/eme...
www.anthropic.com/research/eme...
Reposted by Peter Swimm
Insurers are now seeking to exclude AI from corporate policies because the risks are too unpredictable and liability is unclear. This is a signal: we cannot rely on insurance markets or tort law to protect against AI harms. We need actual governance and policy.
Insurers AIG, Great American, and WR Berkley seek permission to limit liability from AI agents and chatbots and to retreat from coverage of AI risk as multibillion-dollar claims mount
www.ft.com/content/abfe...
www.ft.com/content/abfe...
Insurers retreat from AI cover as risk of multibillion-dollar claims mounts
AIG, Great American and WR Berkley seek permission to limit liability from AI agents and chatbots
www.ft.com
November 24, 2025 at 12:19 PM
Insurers are now seeking to exclude AI from corporate policies because the risks are too unpredictable and liability is unclear. This is a signal: we cannot rely on insurance markets or tort law to protect against AI harms. We need actual governance and policy.
hello god it’s me, charlie brown
November 24, 2025 at 5:09 AM
hello god it’s me, charlie brown
“Don’t let the millions who gave their lives for conflict minerals have done it for nothing.”
www.cnbc.com/2025/11/23/h... "please, buy new phone. the economy is dying" is an incredible gaslight
Americans are holding onto devices longer than ever and it's costing economy
Americans are holding onto devices longer than ever before, and while it may be consumer smart, it comes at a cost to work productivity and the U.S. economy.
www.cnbc.com
November 24, 2025 at 5:06 AM
“Don’t let the millions who gave their lives for conflict minerals have done it for nothing.”
Reposted by Peter Swimm
Officers who can’t handle public scrutiny should get a job that doesn’t involve publicly abducting people.
Immigration agents claim routine reporting violates federal law
Officers who can’t handle public scrutiny should get a job that doesn’t involve publicly abducting people.
freedom.press
November 23, 2025 at 7:50 PM
Officers who can’t handle public scrutiny should get a job that doesn’t involve publicly abducting people.
Reposted by Peter Swimm
Literally EVERY parent should ban Roblox
Every parent whose career has to do with online safety also bans Roblox.
Every game designer parent I know bans Roblox for their kids
November 22, 2025 at 8:47 AM
Literally EVERY parent should ban Roblox
Reposted by Peter Swimm
People who adore the LLM-machine, that by design has no concern for truth, must have no concern for truth
November 23, 2025 at 4:15 PM
People who adore the LLM-machine, that by design has no concern for truth, must have no concern for truth
Reposted by Peter Swimm
You might have heard about NPS. For those of you who don't, I'll briefly explain over the loud groans and boos from the rest of the audience.
NPS is this thing.
Everything about it is stupid but we're going to focus on one way in which it is stupid: it is a trailing indicator.
NPS is this thing.
Everything about it is stupid but we're going to focus on one way in which it is stupid: it is a trailing indicator.
May 30, 2025 at 3:41 PM
You might have heard about NPS. For those of you who don't, I'll briefly explain over the loud groans and boos from the rest of the audience.
NPS is this thing.
Everything about it is stupid but we're going to focus on one way in which it is stupid: it is a trailing indicator.
NPS is this thing.
Everything about it is stupid but we're going to focus on one way in which it is stupid: it is a trailing indicator.
Reposted by Peter Swimm
It is very hard to overstate the evil of the fossil fuel industry.
When people finally do figure it out, it may be too late to undo the damage.
When people finally do figure it out, it may be too late to undo the damage.
U.N. climate talks fizzle out 10 years after Paris accord
Nearly 200 countries at the U.N. climate summit reached a deal that didn’t include a road map to curtail use of fossil fuels, the main driver of climate change.
www.washingtonpost.com
November 23, 2025 at 2:14 PM
It is very hard to overstate the evil of the fossil fuel industry.
When people finally do figure it out, it may be too late to undo the damage.
When people finally do figure it out, it may be too late to undo the damage.
Reposted by Peter Swimm
“Until its A.I. can accomplish some incredible feat — say, generating a cure for cancer — success is partly defined by turning ChatGPT into a lucrative business. That means continually increasing how many people use and pay for it.”
What OpenAI Did When ChatGPT Users Lost Touch With Reality
www.nytimes.com
November 23, 2025 at 12:27 PM
“Until its A.I. can accomplish some incredible feat — say, generating a cure for cancer — success is partly defined by turning ChatGPT into a lucrative business. That means continually increasing how many people use and pay for it.”
Reposted by Peter Swimm
I think the shortest version of why LLMs are an anti education technology is that education is fundamentally about making shared context to understand ourselves, each other, and the world
And by design LLMs destroy shared context
And by design LLMs destroy shared context
November 23, 2025 at 12:57 PM
I think the shortest version of why LLMs are an anti education technology is that education is fundamentally about making shared context to understand ourselves, each other, and the world
And by design LLMs destroy shared context
And by design LLMs destroy shared context
Reposted by Peter Swimm
Reposted by Peter Swimm
what’s funny (it’s not funny) is how many indie journalists, who—with next to none of the infrastructure propping up the zz’s of the business—not only report incisively on political power but do so without leaving years of embarrassing receipts about their relationships with subjects and sources
November 22, 2025 at 4:42 PM
what’s funny (it’s not funny) is how many indie journalists, who—with next to none of the infrastructure propping up the zz’s of the business—not only report incisively on political power but do so without leaving years of embarrassing receipts about their relationships with subjects and sources
Reposted by Peter Swimm
I spoke with people role-playing ICE raids in video games to help educate people on their rights and what to do in real-life situations. Tonight, they're holding an event in Fortnite www.wired.com/story/activi...
November 21, 2025 at 10:16 PM
I spoke with people role-playing ICE raids in video games to help educate people on their rights and what to do in real-life situations. Tonight, they're holding an event in Fortnite www.wired.com/story/activi...
Reposted by Peter Swimm
I've been trying to express exactly this discomfort with modern interface design and especially LLM design and this lady nailed it back in *1993*
("A Small Matter of Programming", Bonnie A. Nardi)
("A Small Matter of Programming", Bonnie A. Nardi)
November 13, 2025 at 3:31 AM
I've been trying to express exactly this discomfort with modern interface design and especially LLM design and this lady nailed it back in *1993*
("A Small Matter of Programming", Bonnie A. Nardi)
("A Small Matter of Programming", Bonnie A. Nardi)
Reposted by Peter Swimm
Srsly, as an academic: this field cites Schmidhuber & LeCun the way biologists cite Darwin. They think they're inventing intelligence from first principles. But their lingo is pidgin cog neuro, they can't define intelligence, & their "theories" are software engineering
dl.acm.org/doi/full/10....
dl.acm.org/doi/full/10....
November 21, 2025 at 6:01 PM
Srsly, as an academic: this field cites Schmidhuber & LeCun the way biologists cite Darwin. They think they're inventing intelligence from first principles. But their lingo is pidgin cog neuro, they can't define intelligence, & their "theories" are software engineering
dl.acm.org/doi/full/10....
dl.acm.org/doi/full/10....