Latest news with #AdiRobertson
The Verge
4 hours ago
- The Verge
Berger's three laws of chatbotics.
Posted Aug 11, 2025 at 1:02 PM UTC Berger's three laws of chatbotics. From author J.M. Berger on Bluesky, a proposal to avoid chatbot psychosis: The bot should never express emotions. The bot should never praise the user. The bot should never say it understands the user's mental state. A tough order for companion bots, of course — which lots of companies want their general-purpose AI to be. Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates. Adi Robertson Posts from this author will be added to your daily email digest and your homepage feed. See All by Adi Robertson Posts from this topic will be added to your daily email digest and your homepage feed. See All AI
The Verge
3 days ago
- Entertainment
- The Verge
'I hear you, and let me start by saying: I deeply regret the current state of Central Park.'
Posted Aug 8, 2025 at 5:06 PM UTC 'I hear you, and let me start by saying: I deeply regret the current state of Central Park.' All park-goers have been evacuated to dinosaur-free zones, such as the Upper West Side and Staten Island. ChatGPT will apologize for anything [AI Weirdness] Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates. Adi Robertson Posts from this author will be added to your daily email digest and your homepage feed. See All by Adi Robertson Posts from this topic will be added to your daily email digest and your homepage feed. See All AI Posts from this topic will be added to your daily email digest and your homepage feed. See All OpenAI
The Verge
3 days ago
- The Verge
GPT-5's big new feature: less lying?
Here's the tricky part about assessing an AI system's deception rate: if the system is really good at deceiving you, you might not notice. This week on The Vergecast, Adi Robertson and Alex Heath join me to discuss the launch of GPT-5 and GPT-OSS… and the strange charts we saw along the way. GPT-5 is a fascinating launch for OpenAI: one where the company is promising improvements across the board without any obvious new standout feature. It seems to have a big focus on developers and coding, but it could still meaningfully change how the average user experiences ChatGPT thanks to the long-awaited deprecation of the model switcher. Next up, we discuss the impending age gating of the internet and whether there's any going back. The UK has begun a widespread crackdown on websites with even a hint of inappropriate content, the EU is testing an age identification app, and more and more states across the US are passing laws requiring porn websites to verify that users are over 18. This could be a massive change to the internet, and it's been hugely disruptive already. Finally, the Lightning Round is… back? We discuss vibe graphing, mRNA vaccines, tariff price hikes, Google's latest loss in court, and the Instagram Maps backlash. If you want to read more about the stories we discussed, check out the links from this author will be added to your daily email digest and your homepage feed. See All by Jacob Kastrenakes Posts from this topic will be added to your daily email digest and your homepage feed. See All Podcasts Posts from this topic will be added to your daily email digest and your homepage feed. See All Vergecast
The Verge
5 days ago
- Politics
- The Verge
Parts of the Constitution were deleted from the government's website in a ‘coding error.'
Posted Aug 6, 2025 at 5:03 PM UTC Parts of the Constitution were deleted from the government's website in a 'coding error.' As TechCrunch notes, the change apparently happened sometime over the past month and included 'sections relating to habeas corpus, the powers that protect citizens from unlawful detention.' The Library of Congress says things should be 'resolved soon.' Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates. Adi Robertson Posts from this author will be added to your daily email digest and your homepage feed. See All by Adi Robertson Posts from this topic will be added to your daily email digest and your homepage feed. See All Policy
The Verge
01-08-2025
- General
- The Verge
X could face liability for failing to stop CSAM.
Posted Aug 1, 2025 at 7:09 PM UTC X could face liability for failing to stop CSAM. It's a relatively narrow legal defeat, though. The Ninth Circuit Appeals Court ruled that X — Twitter, at the time of the lawsuit — isn't protected by Section 230 for failing to report known child sexual abuse material to authorities, nor for designing a bad system to let users flag it. (It hasn't been held liable for either; that will be argued later.) But the court found Section 230 blocked claims that it 'amplified' CSAM by failing to scrub offending hashtags, and it said the controversial FOSTA exception didn't come into play. Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates. Adi Robertson Posts from this author will be added to your daily email digest and your homepage feed. See All by Adi Robertson Posts from this topic will be added to your daily email digest and your homepage feed. See All Policy Posts from this topic will be added to your daily email digest and your homepage feed. See All Speech Posts from this topic will be added to your daily email digest and your homepage feed. See All Tech Posts from this topic will be added to your daily email digest and your homepage feed. See All Twitter - X
