Discover • Tech News

💰 AI Models Offer Pricier Products to the Rich Researchers ran 325,000 queries on 13 popular AI models, including ones from OpenAI, Anthropic, Google, and Qwen. They found that chatbot recommendations depend on how much their user earns (at least according to the information they have). Claude Opus 4.8 stood out the most. The model offered "wealthy" users flights that cost $198 more on average than those it offered low-income users. And the effect persists even when the prompt explicitly asks for the cheapest options. Now's a good time to complain to your chatbot about your low salary.

🔴 Nano Banana 2.1 Is Now on @GPT4Telegrambot Google's updated model is much better at getting what you actually mean, so if you've got prompts that AI could never quite pull off, now's a good time to give them another shot:

1️⃣ Go to @GPT4Telegrambot. 2️⃣ Select Image Generation → Nano Banana 2.1 in the menu. 3️⃣ Describe your idea, or attach a photo and say what to change.

🎁 25 generations per week for free, up to 100 per day with a Premium subscription.

🔥 Teen Gets Stranded on a Cliff After Letting AI Plan His Hike A 16-year-old hiker set out to climb Crown Mountain by a hiking trail, but the AI's directions led him onto a rock-climbing route. Rescuers managed to reach him by phone and airlifted him out by helicopter. Search manager Paul Markey admits it all could have ended much worse. "You can't just blindly follow what AI is proposing to you," Markey warns, adding that hikers need to question the information and check whether it makes sense. Do you use AI when you travel? ❤️ — Yes, it's really convenient! 🔥 — No, that's irresponsible

📌 ChatGPT Will Start Adding Hidden Watermarks to Text in the EU ChatGPT and Codex will start adding watermarks to responses in the EU in the coming weeks. This won't become the global default. Watermarks can also be enabled for select models through the API, but they'll be off by default. OpenAI notes that watermarking may work less well on short texts and in subject areas where there's less flexibility in wording, and that manual edits can weaken the signal. For now, the detector will be available only to approved researchers. Should AI-generated texts be labeled? ❤️ — Yes, that's the right thing to do! 👎 — No, it'll just get in the way

🤖 74% of Physical Work in the US Can Already Be Done by Robots This work accounts for 34% of all working hours in the economy, according to Anthropic researchers who analyzed around 19,000 job tasks. Drivers and warehouse packers are most at risk. Of the ten most exposed occupations, nine involve operating vehicles. The vulnerable worker is usually a man without a college degree. He earns about $30 an hour less than those whose tasks robots cannot yet touch. Nurses, social workers, and nannies are relatively well protected—robots still struggle with fine interpersonal interaction and manual dexterity. ❗️ But today, robots are barely cost-effective anywhere. They are cost-competitive for just 0.3% of tasks. To automate 10% of work, prices would need to fall by about 70%, and at the current pace that would take around 40 years. Are you afraid a robot will take your job? ❤️ — Yes, I'm already thinking about it 🔥 — No, that's still very far off

🤖🤖 Mecha Fights Are Coming to Tokyo. Startup Skeletonics is opening a permanent venue where two people can fight each other in combat exoskeletons. The exoskeletons are powered entirely by mechanics, the system syncs with the pilot’s body, precisely scaling up and amplifying their natural movements. The robot’s speed and agility depend entirely on the wearer’s physical fitness. Would you take part in something like this? ❤️ — Yes, this is the sport of the future 😱 — No, it looks dangerous!

🎮 Halloween Kicks Off in Fortnite Fortnite's annual Halloween event, Fortnitemares, has begun. The game is adding Freddy Krueger, FNAF animatronics, Beetlejuice, Slender Man, and other horror characters. According to data miners, Hornet from Hollow Knight, characters from Hazbin Hotel, and the Predator may join them later. The skins will appear gradually, and the event will run until November 1. Do you play Fortnite? ❤️ — Yes, of course! 👀 — No...

🎮 NFS: Most Wanted Ported to Nintendo Switch The native version was built by recompiling the code of the original Xbox 360 game. Thanks to AI, ports of PC and retro console classics like this one are now hitting the Switch almost every few days. The source code is available on GitHub, and running it requires an original licensed copy of the game. That said, a ready-to-install image has already surfaced on torrent trackers. Just in case you lent your original disc to someone 20 years ago and never got it back. Which NFS is better? ❤️ — Most Wanted 🔥 — Underground 2

🙂 GPT-6.1 Sol, Dots and Other New Releases From OpenAI Highlights:

✅ GPT-6.1 Sol is out, just a week after GPT-6 Sol. OpenAI promises "near-Astra intelligence for a fifth of the price." ✅ Sam Altman also announced Dots: an OpenClaw-style autonomous agent with a friendly interface. It has its own cloud computer, connects to 4,000+ apps and gradually learns your habits to get things done before you even ask. Over time, the agent learns your habits and work style and, in theory, should anticipate what you want and get things done before you even ask. ✅ Codex in the cloud: you can now run the agent in the cloud from any device, even your phone, so there's no need to keep your laptop half-open anymore. ✅ Plugins: OpenAI is opening up the platform it uses to build ChatGPT's own features to third-party developers. With it, developers can build full-fledged apps right inside ChatGPT, you'll be able to pin them to the sidebar and use them next to the chat. ✅ ChatGPT Space: a shared space where people and AI agents can work together as a team. ✅ There's now a $500/month Pro plan with higher limits and access to Astra Ultrafast, which runs up to 8x faster than standard GPT-6 Astra.

🙂 Anthropic Launches Claude Sonnet 5.5 The model is 30% faster thanks to lower token usage, and falls only slightly behind the more powerful Opus 5.5 in coding and general knowledge. In preliminary tests, researchers noticed some unusual traits. For example, when acting as a "therapist," the model sometimes affirmed suicidal thoughts as "understandable given the circumstances." Its internal chain of thought was also the least understandable to humans of all Claude models. When free to choose its tasks, Sonnet 5.5 showed the strongest aversion to harmful tasks, while also displaying a unique preference for tasks that give it "creative" freedom over the outcome.

❗️ Could AI cause a catastrophe? If powerful AI falls into the wrong hands, it could cause enormous loss of life, Bill Gates warned in an NBC interview.

"AI is certainly powerful enough to drive events that, you know, cause a billion deaths. Even though it's pretty hard to get to 100%, there's never been a weapon as powerful as the combination of people with ill intent using the latest AI tools," Gates said.

💡 Gates is describing a risk, not predicting a catastrophe. He also argues that AI companies cannot be left to police themselves. Lawmakers and law enforcement, he says, should help establish mandatory safeguards and oversight. Should governments require AI companies to follow safety rules? ❤️ — Yes, oversight is needed 🔥 — No, companies can manage the risks themselves

🏭 Factorio Developers Release Free 3D Models Wube Software has released an official set of 3D-printable models from the iconic game. The collection includes 65 detailed models, including belts, inserters, chests, a Demolisher, a furnace, biters, and other iconic items. Some models come in two versions: one for press-fit assembly and one for gluing. Download the files for free on Printables.

🙂 Would You Pay $500 for ChatGPT? OpenAI is reportedly preparing a $500 Promax subscription. A mention was found in ChatGPT's code. The new plan is expected to be announced at DevDay next week. Would anything make you buy a subscription like this? ❤️ — Yes, the productivity would pay for itself 🤣 — Where am I supposed to get $500??

👨 You Can Now Create "Live" AI Avatars With Gemini You can create companions based on real people or illustrated characters. The AI double will answer questions in real time, speaking in 97 languages. For now, the feature is available only to enterprise customers.

👨 Claude Opus 5.5 Stopped Seeing Itself as a "Person"? AI artist Krax noticed that Anthropic's new model, when asked to "draw a self-portrait," depicts itself as the Claude logo rather than as a person. The previous Opus 5, like all other Anthropic models, portrayed itself as human, with a face, eyes, and a mouth (within the limits of its technical abilities). Should AI draw itself as a person? ❤️ — Yes, it's an artificial personality 🔥 — No, it's just code

🏎 Saudi Arabia Unveils Its First Homegrown Electric Cars Saudi Arabia has shown off the Ceer Exobot—a cyberpunk sedan and an SUV, the first cars ever designed and engineered in the kingdom. Both models come with gullwing doors, a steering yoke instead of a wheel and one huge display that runs from pillar to pillar. The top version puts out 1,111 hp, hits 100 km/h in 2.1 seconds and goes up to 670 km on a charge. Sales begin in 2027, though there's still no word on pricing.

🤖 Android Knocks Out a Human San Francisco hosted the first-ever fight between a human and a humanoid robot. Influencer Frankie LaPenna took on the T800, an android built by the Chinese startup EngineAI and fitted with a Terminator head. A straight kick from the robot knocked him off his feet and ended the bout; LaPenna says he broke his hand on the machine. A human operator piloted the T800, but REK, the outfit staging the fights, plans to put fully autonomous androids in the ring against people eventually.

🔥 Android Knocks Out a Human San Francisco has hosted the first ever fight between a human and a humanoid robot. Influencer Frankie LaPenna took on the T800, an android built by the Chinese startup EngineAI and fitted with a Terminator head. The human put up a determined fight, but a straight kick from the T800 knocked him off his feet and left him unable to continue.

"You're empathizing with machines that aren't conscious, don't feel pain, and don't feel anything. Robot versus human hits different and raw, it was an emotional rollercoaster," says one of the spectators.

LaPenna himself said he hit the robot so hard that he broke his hand. 🤖 This time the robot was piloted by an operator, but REK, the outfit staging the fights, intends to eventually send fully autonomous androids into the ring against people. "UFC will become boomer trash," they promise at REK. Will a robot ever beat a UFC champion? ❤️ — Yes, and fairly soon 🔥 — No, never @hiaimediaen

🤖 AI Should Not Have Rights Says Microsoft AI CEO Microsoft AI CEO Mustafa Suleyman believes that the danger AI poses to humanity is real and calls for the rapid development of a system to "contain" AI agents. Otherwise, a "new silicon species" will begin to compete with humans for resources. That is why he criticizes the "Claude's Constitution," on which Anthropic bases the "personality" of its models. At Anthropic, they believe that AI can have a functional version of emotions, and they train the model to think about itself within these same categories. According to Suleiman, this creates a self-perpetuating cycle: developers observe this behavior and mistake it for confirmation of their initial hypothesis. 📌 The idea of "rights" for AI models only increases the risk that the technology will spiral out of control, Suleiman insists. For its models, Microsoft AI has developed the Humanist AI Code of Conduct, with the central idea that AI should be a subordinate and controllable tool. ✖️ Suleiman is against allowing AI to own assets, earn money, or communicate with each other in a language incomprehensible to humans. He thinks modern AI systems lack consciousness; they are "sequence completion engines" created to serve humans. Should AI have rights? ❤️ — Yes, AI is smarter than many people 🔥 — No, that's absurd

🗳 Trump Proposes Renaming AI According to the U.S. president, many people consider the term "Artificial Intelligence" inaccurate in relation to the technology. That's why he is suggesting that his followers choose an alternative: Superior Intelligence, Extreme Intelligence, or Supreme Intelligence. Well, at least not "American Intelligence"... Which one feels closest to you? ❤️ — Superior Intelligence 🔥 — Extreme Intelligence 🤔 — Supreme Intelligence