OpenAI tokenmaxxing as Meta tracks AI usage

Bluesky recently introduced Attie, a new AI application designed to help users create personalized content feeds. This standalone app, developed by former CEO Jay Graber's team, allows users to input natural language prompts to find relevant posts across Bluesky and other compatible networks. The long-term vision is to integrate Attie's custom feed capabilities directly into Bluesky and other applications built on the AT Protocol, eventually enabling users to "vibe code" entire apps without traditional coding knowledge.

The increasing prevalence of AI chatbots also brings important discussions about safety, especially for young people. Educators and law enforcement officials emphasize the need for open conversations with adolescents regarding potential risks, particularly following a lawsuit filed against OpenAI related to a teen's suicide. Meanwhile, scammers are actively exploiting AI, with over half of spam emails and 82% of phishing emails now AI-generated. Deepfakes, including AI-created images and audio, are also increasingly used in imposter scams, making verification crucial.

In personal relationships, a trend dubbed the "Claude-gap" describes couples where one partner is deeply immersed in AI "vibe coding" while the other remains uninterested. On the corporate front, "tokenmaxxing" is emerging at major tech companies like Meta and OpenAI, where internal dashboards track employee AI usage as a performance metric. This practice aims to automate workflows, with some engineers allocating significant funds to AI agents to automate tasks, which in turn boosts revenue projections for AI companies.

Google's Jeff Dean addresses concerns about AI running out of training data, suggesting that vast amounts of video, audio, robotics, and autonomous vehicle data remain largely untapped. He believes computing power, rather than data availability, may become the primary constraint for AI advancement. In a practical application, a Florida man successfully sold his home for $954,800, exceeding agent estimates by $100,000, by using ChatGPT for advice on value-increasing upgrades.

However, former Facebook leader Justin Rosenstein warns that AI is following a path similar to social media, prioritizing profit over public good. He advocates for public control of AI through citizens' assemblies to ensure it serves humanity's best interests. Additionally, a new essay collection explores the intersection of artificial intelligence, computer vision, the surveillance state, and magic, reflecting broader societal engagement with AI's implications.

Key Takeaways

  • Bluesky launched Attie, an AI app by Jay Graber's team, for personalized feeds and future "vibe coding" of AT Protocol apps.
  • Educators urge open talks with youth about AI chatbot risks, citing an OpenAI lawsuit concerning a teen's suicide.
  • Over 50% of spam and 82% of phishing emails are AI-generated, with deepfakes increasingly used in scams.
  • Google's Jeff Dean states AI data concerns are exaggerated, pointing to untapped video, audio, robotics, and autonomous vehicle data.
  • The "Claude-gap" trend highlights relationships where one partner is engrossed in AI "vibe coding" while the other is not.
  • A Florida man used ChatGPT advice to sell his home for $954,800, exceeding estimates by $100,000.
  • "Tokenmaxxing" is a trend at Meta and OpenAI where employee AI usage is tracked as a performance metric.
  • Justin Rosenstein, a former Facebook leader, warns AI prioritizes profit over public good and calls for public control via citizens' assemblies.
  • Engineers are funding AI agents to automate tasks, increasing revenue projections for AI companies.
  • A new essay collection explores AI, computer vision, the surveillance state, and magic.

Bluesky launches Attie, an AI app for personalized feeds

Bluesky has released a new standalone app called Attie, which uses AI to help users create personalized content feeds. Users can input prompts in natural language, and Attie will find relevant posts across Bluesky and other compatible networks. Former CEO Jay Graber's new team developed this people-focused AI product. Initially, Attie will be a separate app, but the goal is to integrate its custom feed capabilities into Bluesky and other applications built on the AT Protocol. Attendees at the 2026 Bluesky Atmosphere conference were able to beta test Attie, which aims to allow users to 'vibe code' their own apps.

Bluesky's new app Attie uses AI to customize your feed

Bluesky has launched Attie, a new AI assistant designed to create custom content feeds based on user prompts. Users can ask Attie for specific topics, like 'posts about folklore, mythology, and traditional music, especially Celtic traditions.' While these custom feeds will initially be in a standalone Attie app, the plan is to make them available in Bluesky and other AT Protocol apps. Eventually, Attie will allow users to 'vibe code' entire applications on top of the AT Protocol, making app creation accessible without coding experience.

Educators urge open talks with kids about AI chatbot risks

Local educators and law enforcement officials emphasize the importance of open conversations with young people about AI chatbots due to their increasing prevalence. They stress that strong relationships are key to protecting adolescents from potential dangers, especially after a lawsuit was filed against OpenAI concerning a teen's suicide. While regulations and safeguards are being developed, clear communication between families and schools about AI usage is crucial. Experts advise parents to educate themselves about different AI tools, encourage critical thinking about AI-generated content, and set boundaries rather than outright bans.

The 'Claude-gap' trend: Couples navigate AI coding obsession

A growing trend called the 'Claude-gap' describes relationships where one partner is deeply involved in AI 'vibe coding' while the other is uninterested. Couples like Brian Deagan and Leslie Kemelgor, and Kendra Ramirez and her husband, share experiences of navigating these differing interests. Some couples establish boundaries, like no AI coding during family time, to maintain balance. While AI coding can be captivating, partners who are not involved often feel left out or struggle to understand the technical details.

Google's Jeff Dean: AI still has vast untapped data sources

Google's Jeff Dean believes concerns about running out of data to train AI models are exaggerated. He highlighted that significant amounts of video and audio data, along with data from robotics and autonomous vehicles, are not yet fully utilized for AI training. Dean also discussed the potential of synthetic data and techniques like data augmentation and regularization to improve AI models. He suggested that computing power, rather than data availability, might become the primary constraint for advancing AI.

Scammers exploit AI: How to protect yourself

Scammers are enthusiastically using artificial intelligence, with over half of spam emails and 82% of phishing emails now AI-generated. Deepfakes, including AI-created images, videos, and audio of public figures or even ordinary people, are becoming increasingly common in scams. These AI tools can be used for imposter scams, fake advertisements, and more, making it difficult for humans to distinguish reality from fabrication. Experts advise using common sense, stopping to think before acting, talking to trusted individuals, and verifying information to combat AI-powered scams.

Florida man sells home for $954,800 using ChatGPT advice

A Florida man successfully sold his family home for $954,800, exceeding agent estimates by $100,000, by utilizing ChatGPT. The AI tool provided recommendations for specific upgrades that were believed to increase the property's value. This case highlights how AI can assist in real estate transactions by offering data-driven advice on maximizing sale prices.

Tokenmaxxing: The AI status game impacting work

Tokenmaxxing is a new trend where companies track employee AI usage as a performance metric, turning AI consumption into a visible status game. This practice has spread to major tech companies like Meta and OpenAI, with internal dashboards monitoring 'tokens' used by employees. While some see it as a way to automate workflows and boost productivity, critics argue it encourages a race for perceived productivity rather than actual achievement. Some engineers are even allocating significant funds to AI agents to automate tasks, leading to increased revenue projections for AI companies.

New book explores AI, psyops, and magic

This week's recommendations include a new essay collection that delves into artificial intelligence, computer vision, the surveillance state, UFO photos, psyops, and magic. The roundup also features articles on bureaucracy, 'Killers of the Flower Moon,' and the state of Western intellectual production. Additionally, there's an entertaining video essay about video stores in Hollywood films and other media like a video of a mouse on a blueberry and music tracks.

Facebook co-founder warns AI needs public control

Justin Rosenstein, a former Facebook leader, warns that artificial intelligence is heading down a similar path of prioritizing profit over public good, much like social media. He argues that AI is already governing many aspects of our lives without public input and that companies are rushing to deploy it despite safety concerns. Rosenstein advocates for public control of AI through citizens' assemblies, similar to how Ireland has addressed complex social issues. He believes this democratic approach is necessary to ensure AI serves humanity's best interests, not just corporate profits.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI applications AI ethics AI safety AI trends AI tools AI-generated content AI regulation AI in real estate AI in social media AI in work AI models AI development AI assistants AI personalization AI data AI scams AI coding AI education AI for business AI for consumers AT Protocol Bluesky ChatGPT Claude Deepfakes OpenAI Tokenmaxxing Vibe coding

Comments

Loading...