Dinner with Sam Altman

Good morning, AI enthusiasts. Despite just launching GPT-5, Sam Altman just admitted what many suspected — OpenAI has higher AI models sitting on the shelf that we won’t even access yet as a result of compute constraints.

But that revelation was just the appetizer in a dinner conversation that covered every thing from trillion-dollar infrastructure plans, Jony Ive’s device and “recent computing paradigm”, and even a possible Google Chrome acquisition.

In today’s AI rundown:

• Altman details OpenAI’s trillion-dollar roadmap

• Anthropic gives Claude the ability to ‘hang up’

• Automate meeting prep with ChatGPT

• GPT-5 blows past doctors on medical exams

• 4 recent AI tools, community workflows, and more

LATEST DEVELOPMENTS

OPENAI

🍽️ Altman details OpenAI’s trillion-dollar roadmap

The Rundown: OpenAI hosted reporters from outlets including TechCrunch and The Verge over dinner, speaking on topics from GPT-5’s reception to the corporate’s plans for social media, consumer hardware, and a possible Chrome acquisition.

The main points:

• Altman said he “legitimately just thought we screwed that up” on 4o’s removal, with GPT-5 focused on warmer responses while not being sycophantic.

• He revealed OAI has higher models they will’t offer as a result of compute constraints, saying they’ll spend “trillions” on data centers within the near future.

• Altman acknowledged parallels between the AI frenzy and the dot-com bubble, calling valuations “insane” but saying the tech justifies massive investments.

• He also commented on Perplexity’s Google Chrome bid, saying OpenAI should “take a have a look at it” if the browser is forced to be sold in the present legal battle.

• The CEO reiterated the corporate’s device with Jony Ive will probably be “well worth the wait,” confidently saying, “you don’t get a brand new computing paradigm fairly often”.

Why it matters: Despite OpenAI’s astronomical rise and trillion-dollar ambitions, these candid moments offer the AI world something rare — each a glance behind the scenes of the buzziest company on the planet and a fly-on-the-wall glimpse of the longer term through the eyes of one among tech’s strongest (and polarizing) figures.

TOGETHER WITH CONCIERGE

👋 Your brand’s AI answer engine

The Rundown: Today’s SaaS buyers use AI each day to reply their questions, and haven’t any patience for a scavenger hunt. Concierge is a Perplexity-style answer engine, trained on your organization, that lives in your website and delivers accurate, personalized responses to ultra-specific questions.

Modern brands use Concierge to:

• Handle any buyer query (regardless of how technical) with advanced RAG in your sources & media

• Control and visibility over every conversation, with guardrails and sentiment evaluation

• Construct trust with website visitors before they’re willing to commit to a demo

Try Concierge to show every query right into a conversation — and each conversation into revenue.

ANTHROPIC

🛑 Anthropic gives Claude the ability to ‘hang up’

The Rundown: Anthropic just equipped Claude Opus 4 and 4.1 with the power to finish chats believed to be harmful/abusive as a part of the corporate’s research on model wellness, marking one among the primary AI welfare deployments in consumer chatbots.

The main points:

• The tip chat feature will trigger after Claude’s redirections and productive engagement fails on content requested about minors, terrorism, or violence.

• Testing revealed that Opus 4 exhibited distress patterns when processing harmful requests, voluntarily terminating simulated abusive interactions.

• Despite the “hang up,” users still retain full account access and might immediately start fresh conversations or edit previous messages.

• Anthropic has also programmed safeguards stopping ending messages when users show signs of self-harm risk or imminent danger to others.

Why it matters: Anthropic is one among the few labs putting serious time into model welfare — and while no person truly knows where things stand with AI systems because it pertains to consciousness, we may look back on this research as vital first steps for a phenomenon that doesn’t have a transparent precedent or roadmap.

AI TRAINING

📝 Automate meeting prep with ChatGPT

The Rundown: Use ChatGPT’s Gmail, Google Calendar, and Google Drive connectors to robotically research attendees, pull email context, and gather relevant files for comprehensive pre-meeting briefs.

Step-by-step:

1. Click your profile → Connectors in ChatGPT and enable Gmail, Google Calendar, and Google Drive (requires Plus/Pro account)

2. Prompt: “Analyze my next meeting and supply: attendee backgrounds from email history, meeting context, relevant Drive files, and key discussion points”

3. ChatGPT robotically searches all connected sources and returns an in depth transient with agenda items, project updates, and suggested documents

4. Arrange automation: “Create a each day 9 AM task to prep all my meetings for the day”

Pro tip: For top-stakes meetings, add “Discover potential concerns based on our email history” to anticipate objections and prepare stronger responses.

PRESENTED BY FIN

🔮 The long run of AI customer support is here

The Rundown: Pioneer by Fin is a summit that brings together AI customer support leaders to explore how AI Agents are transforming support. Learn from experts at Anthropic, BCG, Rocket Money, and more about customizing, testing, and scaling Fin across every channel.

Register today to affix Pioneer online or in NYC, where you’ll:

• Hear Intercom leadership’s vision for the longer term of customer support

• Learn the way customer support leaders are adopting AI-first strategies

• See the most recent Fin product launches

Register now.

AI RESEARCH

🏥 GPT-5 blows past doctors on medical exams

The Rundown: OpenAI’s GPT-5 posted impressive results on medical reasoning benchmarks, surpassing each GPT-4o and human medical professionals by substantial margins across diagnostic and multimodal tasks in a brand new study from Emory University.

The main points:

• The model achieved 95.84% accuracy on MedQA’s clinical questions, jumping 4.8 percentage points over GPT-4o’s previous best.

• GPT-5 scored 70% on multimodal medical reasoning tasks that mix patient histories with imaging, gaining nearly 30 points over GPT-4o.

• The system also exceeded pre-licensed medical professionals by 24% on reasoning and 29% on understanding in expert-level tests.

• GPT-5 showed sophisticated diagnostic abilities on complex cases, appropriately ID’ing rare conditions like Boerhaave syndrome from lab values and CT scans.

Why it matters: The shift from GPT-4o’s near-human performance to GPT-5’s superiority over medical professionals shows we’re approaching some extent where physicians NOT using AI in clinical settings could possibly be considered malpractice (H/T Dr. Derya Unutmaz). Plus, the gap is barely heading in a single direction as intelligence scales.

QUICK HITS

🛠️ Trending AI Tools

• 📸 Imagen 4 – Google’s SOTA image generation model, now generally available

• 🎬 Higgsfield Product-to-Video – Add product placement straight into videos

• ⚙️ Genspark AI Developer – Construct anything without coding knowledge

• 👋 GPT-5 – Recent ‘warmer’ and ‘friendlier’ personality upgrade

📰 All the pieces else in AI today

Meta is reportedly planning one other restructure of its AI divisions, marking the fourth in only six months, with the corporate’s MSL set to be divided into 4 teams.

StepFun AI released NextStep-1, a brand new open-source image generation model that achieves SOTA performance amongst autoregressive models.

Meta FAIR introduced Dinov3, a brand new AI vision foundation model that achieves top performance with no labeled data needed.

The U.S. government rolled out USAi, a platform for federal agencies to utilize AI tools like chatbots, coding models, and more in a secure environment.

OpenAI’s GPT-5 had essentially the most success of any model yet in tests playing old Pokémon Game Boy titles, beating Pokémon Red in nearly a 3rd of the steps as o3.

COMMUNITY

🤝 Community AI workflows

Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.

Today’s workflow comes from reader Adrian P. in South Africa:

“I’m within the production/engineering and customarily use AI to assist construct apps that streamline what could be previously laborious and really paper-intensive processes into web-based systems. This helps to save lots of time, increase green footprint, and allows for the tracking and tracing of information at any date in the longer term, which in turn helps with problem solving and resolving customer complaints.”

How are you using AI? Tell us here.

🎓 Highlights: News, Guides & Events

• Read our previous newsletter: Superbugs meet their AI match

• Today’s AI tool guide: Automate meeting prep with ChatGPT connectors

• Watch our last live workshop: Mastering AI agents for customer support

See you soon,

Rowan, Joey, Zach, Shubham, and Jennifer—the humans behind The Rundown