PLUS: Anthropic's $50B investment, a new screen-watching AI, and AI music on Billboard
Good morning
OpenAI has just launched a significant update to ChatGPT with GPT-5.1, which introduces new preset ‘personalities’ to customize its communication style. This release aims to make the AI feel less like a generic tool and more like a tailored assistant.
The update, which also includes adaptive reasoning for faster performance, pushes the AI towards a more human-like, fine-tuned experience. But as AI becomes more personal, where does OpenAI draw the line between a helpful tool and one that fosters unhealthy user attachment?
In today’s Next in AI:
ChatGPT's new preset personalities
Anthropic's $50B infrastructure investment
A new screen-watching AI assistant
AI music tops the Billboard charts
ChatGPT Gets Personal
Next in AI: OpenAI just rolled out GPT-5.1, a major update that makes ChatGPT faster, smarter, and more conversational. The release introduces new preset ‘personalities’ that let you tailor its communication style for any task.
Decoded:
The update adds new personality presets like “Quirky” and “Candid,” aiming to move beyond a one-size-fits-all approach for its 800 million users.
Under the hood, the new model uses adaptive reasoning, automatically spending more computational time on complex problems while delivering faster replies for simple queries.
Paid subscribers get first access, with a gradual rollout to all users, and the previous GPT-5 models will remain available for three months to help everyone adapt.
Why It Matters: This move shows OpenAI is focusing heavily on user experience, making AI feel more like a personal assistant you can fine-tune. The challenge now is to offer this deep personalization without encouraging unhealthy attachments to the technology.
Anthropic's $50B Buildout
Next in AI: Anthropic is pouring a massive $50 billion into American computing infrastructure, building custom data centers in Texas and New York to power the next generation of its AI models. This project helps advance goals in the government's AI Action Plan to maintain American leadership in the field.
Decoded:
This investment tackles a massive energy bottleneck, with Anthropic projecting the U.S. AI sector will need at least 50GW of electric capacity by 2028—roughly twice New York City's peak demand.
The project, built in partnership with data center specialist Fluidstack, is expected to create 800 permanent jobs and 2,400 construction jobs, stimulating the broader tech infrastructure economy.
This domestic buildout is a direct response to global competition, as the U.S. works to keep pace with China, which added over 400 gigawatts of power capacity last year alone.
Why It Matters: This massive capital injection is not just about building data centers; it is about securing the physical supply chain needed for America to lead in the next era of AI. The move signals that the race for AI dominance will be won not just with algorithms, but with concrete, steel, and immense power generation.
The All-Seeing Assistant
Next in AI: A new open-source AI assistant called Seeva can be summoned anywhere on your desktop with a keystroke. It "watches your screen" to provide context-aware help, aiming to eliminate the constant back-and-forth between apps.
Decoded:
Seeva captures your screen with a single click, using Claude's vision API to understand what you're working on in any application.
It supports multiple AI providers, allowing you to connect your own keys for Anthropic, OpenAI, Google Gemini, or even run local models with Ollama.
The tool is built with a privacy-first approach, storing all conversations, settings, and screen captures locally on your device in an SQLite database.
Why It Matters: Seeva directly tackles the flow-breaking problem of context switching that plagues modern workflows. By offering a customizable and private on-demand assistant, it puts powerful, context-aware AI right where you need it, instantly.
AI's Chart Topper
Next in AI: AI-generated music is now so convincing that 97% of listeners can't tell it apart from human-made tracks, according to a new survey. This surge in quality is sending AI artists straight to the top of the Billboard charts.
Decoded:
A recent Deezer survey of 9,000 people found 97% of listeners could not distinguish between music created by humans and tracks generated entirely by AI.
This isn't just a novelty; an AI country song called “Walk My Walk” recently hit No. 1 on a Billboard chart, and AI R&B singer Xania Monet landed a multimillion-dollar record deal.
Artists are using tools like Suno and Udio to rapidly produce music, raising questions about whether AI-generated tracks can ever feel human as they flood streaming platforms.
Why It Matters: The line between human artistry and machine creation in music is becoming nearly invisible. This shift opens up new creative avenues for individuals but also pressures streaming services to address content authenticity and discovery.
AI Pulse
Data suggests OpenAI's inference compute costs are far higher than its reported revenue, showing a potential cash burn of nearly $5.6B in the first two quarters of 2025 alone.
Researchers found that some users are forming deeply committed relationships with AI chatbots, engaging in role-played marriages and even pregnancies to solidify their bonds.
Yoshua Bengio became the first researcher to surpass one million citations on Google Scholar, a milestone that underscores the massive influence and growth of the machine learning field.
Stanford used AI to design entirely new viruses capable of killing bacteria, a breakthrough that raises complex questions about creating novel life forms.