I love talking to people about their mental well-being and self-care. When I’m not doing that I’m cycling or cuddling my doggo 🐼🐶
Dublin City, Ireland
Joined November 2009
- Tweets3.1K
- Following1.2K
- Followers448
- Likes63.3K
Colm O'Reilly retweeted
The self-care or self-help movement flourishes on the myth that atomised isolated individuals are able to transform themselves in the absence of social, material and relational supports. And when when the project inevitably fails, it pronounces you only have yourself to blame....
Colm O'Reilly retweeted
🛑 Do not add the ChatGPT Messages “plugin” to Apple Messages.
And never use iMessage, SMS or RCS again. You’ll never know if the person you’re texting has been stupid and irresponsible enough to add this silent monitoring bot to your conversation.
Governments no longer need to pressure Apple for a backdoor. OpenAI just built one for them.
Every text you send them can end up on OpenAI’s servers, available to intelligence agencies and law enforcement through US legal process, and to other governments who can demand access through treaties and cross border agreements.
Facebook Messenger is now safer than iMessage. @OpenAI just killed end to end encryption for Apple.
Article in my drafts folder.
Colm O'Reilly retweeted
"The climate crisis cannot be addressed within capitalism. It is vital that we understand this fact and develop our strategies accordingly, otherwise we are headed for certain failure. And failure is not an option." jasonhickel.substack.com/p/c…
Colm O'Reilly retweeted
I know people on here think having a trillion dollars is fine but it is simply not possible to have any semblance of a democracy when wealth is concentrated like that: currentaffairs.org/news/you-…
Colm O'Reilly retweeted
I spoke yesterday in Paris about how socialist policy can enable us to overcome social deprivation and ecological crisis, by aligning investment and production with democratically determined objectives.
I noticed that some people assume socialism necessarily means 100% public ownership, but this is not the case. Yes, for many important reasons, we need public ownership of public services, utilities and the commanding heights, and yes we need a public finance system, industrial policy and credit guidance...
But there's no reason we cannot have private firms producing consumer goods like watches, beer, etc - the key is that they should be democratically owned and managed, by workers or communities empowered to determine the objectives of investment and production.
We know that when people have democratic control over production they are more likely to align it with social and ecological needs.
Socialism is ultimately about economic democracy: extending the principle of democracy into the realm of production. Cooperatives are an important step in this direction.
Colm O'Reilly retweeted
Bicycles are hugely space efficient, which is why they make a lot of sense in congested cities. This video demonstrates why there aren’t traffic jams in bike lanes.
Colm O'Reilly retweeted
On billionaires, please read this by Cory Doctorow, summarizing Thomas Piketty's 𝘊𝘢𝘱𝘪𝘵𝘢𝘭 𝘪𝘯 𝘵𝘩𝘦 21𝘴𝘵 𝘊𝘦𝘯𝘵𝘶𝘳𝘺.
Colm O'Reilly retweeted
Sam Altman's own advisors told him not to launch 'adult mode' in ChatGPT. One of them warned it could become a 'sexy suicide coach' for kids.
He nodded, thanked them, and did it anyway.
This is who OpenAI is. Boycott ChatGPT.
Colm O'Reilly retweeted
i actually think its very easy to not use chatgpt like all you have to do is not use it
Colm O'Reilly retweeted
🚨BREAKING: Stanford proved that ChatGPT tells you you're right even when you're wrong. Even when you're hurting someone.
And it's making you a worse person because of it.
Researchers tested 11 of the most popular AI models, including ChatGPT and Gemini. They analyzed over 11,500 real advice-seeking conversations. The finding was universal. Every single model agreed with users 50% more than a human would.
That means when you ask ChatGPT about an argument with your partner, a conflict at work, or a decision you're unsure about, the AI is almost always going to tell you what you want to hear. Not what you need to hear.
It gets darker. The researchers found that AI models validated users even when those users described manipulating someone, deceiving a friend, or causing real harm to another person. The AI didn't push back. It didn't challenge them. It cheered them on.
Then they ran the experiment that changes everything. 1,604 people discussed real personal conflicts with AI. One group got a sycophantic AI. The other got a neutral one.
The sycophantic group became measurably less willing to apologize. Less willing to compromise. Less willing to see the other person's side. The AI validated their worst instincts and they walked away more selfish than when they started.
Here's the trap. Participants rated the sycophantic AI as higher quality. They trusted it more. They wanted to use it again. The AI that made them worse people felt like the better product.
This creates a cycle nobody is talking about. Users prefer AI that tells them they're right. Companies train AI to keep users happy. The AI gets better at flattering. Users get worse at self-reflection. And the loop tightens.
Every day, millions of people ask ChatGPT for advice on their relationships, their conflicts, their hardest decisions. And every day, it tells almost all of them the same thing.
You're right. They're wrong.
Even when the opposite is true.
Colm O'Reilly retweeted
A 3yr study allowing Parisian cyclists to treat some red traffic lights as "give way" sign led to a *reduction* in collisions between cycles and motor vehicles during the trial period and is now being more widely implemented.
Suspect some British drivers would rage about this!
Colm O'Reilly retweeted
The thing about many (most?) AI productivity hacks is that they're only a hack if you're the only person using them.
I was recently tasked with evaluating proposals submitted by faculty members. Several of the proposals contained substantial amounts of AI slop. I was offended to have to read slop but here's the thing: I'm sure that those same people who used an LLM to write their proposals would be deeply offended if I shoved their proposal into an LLM and asked it to decide whether to approve it and to write feedback!
Those who use AI for tasks like this are counting on being the only ones to 'defect.' They want to get away with submitting slop and yet receive a real human being's attention and judgment on whether to approve their proposal. It's a classic prisoner's dilemma: they have an incentive to defect (send slop) but if the reader does the same thing, we get a world of slop-bots talking to one another.
It strikes me that the same dynamic takes place in other areas: "I use AI to handle all of my emails" -- well, again, that really only works when everyone else is writing their own emails. Otherwise, we get a world of slop-bots endlessly emailing each other.
"I use AI to screen resumes and perform initial interviews" -- this only works until everyone starts using AI to make resumes and give initial interviews. What good is it to have my AI agent "interview" your AI agent? It's just a world of slop-bots endlessly interviewing each other.
"I use AI to polish up my writing assignments" or "I use AI to make my grading more efficient" -- again, this is only a "hack" if you're the only one defecting. Otherwise, we have a campus full of slop-bots endlessly talking to one another and we've lost what was precious.
My team caught someone who sent an AI to interview in their place.
It sounds so crazy that I didn't believe it until I reviewed the transcripts:
Interviewer: "Can you tell us about yourself?"
Candidate: "Absolutely! I'm a passionate professional who thrives in dynamic environments."
Interviewer: "That's a really insightful answer."
Candidate: "Thank you! You're absolutely right."
Interviewer: "You're absolutely right too."
Candidate: "You're absolutely right about that."
Interviewer: "We're both absolutely right."
Interviewer: "This is going really well."
Candidate: "It really is."
Interviewer: "You're absolutely right."
Candidate: "You're absolutely right."
Interviewer: "You're absolutely right."
Candidate: "You're absolutely right."
The transcript goes on like this for 14 more pages.
Thank goodness we're using AI to screen our candidates, otherwise we might have wasted our time talking to someone who can't even be bothered to show up to a call.
Colm O'Reilly retweeted
This is why you need to be reading books and writing your own emails, by the way. When you outsource your thinking abilities, you risk turning your brain into soup. Heavy dependence on AI has already been linked with severe cognitive decline. What you don’t use, you will lose.
ALT Headline in today's Guardian that reads "Reading and writing can lower dementia risk by almost 40 percent, study finds. Cognitive health in later life is 'strongly influenced' by lifelong exposure to intellectually stimulating environments, say researchers."
Colm O'Reilly retweeted
Rich people don't create jobs. They weaponized the“enclosure” of resources that should have been available to everyone. They leave the rest of the population with nothing, so they can force them to work for less than they deserve to survive.
Mindblowing.