OpenAI's big reveal

PLUS: MIT studies AI's deception

Welcome, AI enthusiasts.

OpenAI’s big reveal day is here — and the latest rumors include a new ‘Her’ style voice assistant that could put Siri to shame.

But with the company also reportedly nearing a deal with Apple to integrate into iOS, the implications of today’s demo could be massive. Let’s explore…

In today’s AI rundown:

  • Rumor mill: OpenAI’s Monday reveal

  • Anthropic’s new tool automates prompting

  • How to create realistic AI avatar videos

  • MIT studies AI deception capabilities

  • 6 new AI tools & 4 new AI jobs

  • More AI & tech news

Read time: 4 minutes



Image source: The Telegraph

The Rundown: OpenAI is set to demo new features and updates to ChatGPT and GPT-4 today at 10 AM PT, with new speculation including a ‘Her’ style voice assistant with both audio and visual capabilities.

The details:

  • According to The Information, OpenAI’s demo will include a virtual assistant with visual AND audio understanding.

  • The report also claims the new reveal might have the ability to make ‘existing voice assistants like Siri more useful.’

  • CEO Sam Altman shot down rumors of a new search engine competitor and GPT-5, but said the reveal is something that ‘feels like magic’.

  • Additional speculation includes the ability to initiate and receive phone calls inside of ChatGPT.

  • Apple and OpenAI are also reportedly ‘closing in’ on a deal to incorporate ChatGPT into iOS 18.

Why it matters: OpenAI's typical cryptic commentary has continued to kick hype and speculation into overdrive. While it looks like we’ll have to wait a bit longer for OpenAI’s Search Engine ‘Google Killer,‘ we might just get something even more magical.

You can watch the livestream here, kicking off at 10 AM PT. If you can’t make it, we’ll fill you in on all the details in tomorrow’s newsletter.


The Rundown: WorkOS is a complete auth solution that helps B2B SaaS companies quickly become Enterprise Ready. It’s used by hundreds of high-growth startups like Perplexity, Jasper,, Cursor, and more.

With WorkOS, you can:

  • Add foundational auth as well as complex enterprise features like SSO, SCIM provisioning, and RBAC

  • Utilize flexible and easy-to-use APIs that can be integrated in minutes

  • Onboard IT teams to SSO in minutes with the Admin Portal

  • Enjoy User Management that is free up to 1 million MAUs

Get started with WorkOS for free.


Image source: Anthropic

The Rundown: Anthropic just launched a new Prompt Generator tool for its business and API users, helping to automatically craft optimal prompts via natural language when completing tasks with its Claude models.

The details:

  • The generator leverages advanced prompt techniques like chain-of-thought reasoning for more ‘effective, precise, and reliable’ outputs.

  • Console users can also test prompt performance via dynamic variable insertion, optimizing prompts based on various situations.

  • Anthropic released a Prompt Library earlier this year, featuring a range of optimized prompts that users can copy and paste.

Why it matters: While ‘Prompt Engineer’ was a popular term thrown around as a potential future job, the reality is that AI can help simplify the task with optimal prompts that it creates on its own. While Anthropic’s tool is only on the API side for now, it's only a matter of time before similar features make their way to the full consumer side.


The Rundown: HeyGen lets you create personalized videos with lifelike AI-generated avatar clones and unique AI voices.


  1. Head over to HeyGen’s website and sign up for free.

  2. Click where it says ‘AI Studio’ on the left bar of the Dashboard.

  3. Choose a given template by selecting ‘Templates’ or create one from scratch by pressing ‘Create with AI Studio’.

  4. In the Studio, you can add and modify any part of your presentation. For example, you can add text and images, change the background, select an AI avatar and its voice, create a script using AI, and more.

  5. When your video is ready, press “Submit” and check out the final video!


The Rundown:  The AI Conference brings together leading voices from OpenAI, Anthropic, Meta, Google DeepMind, and more to push the boundaries of AI innovation.

Attend the two-day event in San Francisco, CA on Sept. 10-11 and:

  • Engage with 60+ speakers leading the AI revolution

  • Network, collaborate, and co-create with industry pioneers

  • Explore topics including AGI, AI in enterprise, building with AI, and more

Register now and don’t miss the final chance to secure Early Bird pricing— use discount code rundown24 for an extra 10% off.


Image source: MIT

The Rundown: MIT researchers just published a new study detailing AI’s ability to engage in deceptive behaviors like lying, bluffing, and disguising true capabilities in testing.

The details:

  • The analysis found numerous examples of AI deception, such as Meta's Cicero Diplomacy bot telling premeditated lies to players.

  • Other cases included negotiation systems misrepresenting preferences, AI bluffing in poker, and ‘playing dead’ to fool safety tests.

  • Growing deceptive capabilities could influence societal dangers like election tampering, fraud, and general control over AI systems.

Why it matters: The combination of AI’s superhuman persuasiveness and the growing ability to engage in deception are a reminder of a potential dark side to the tech’s advancement. As these qualities emerge, the importance of reliably aligning and better understanding model behavior only grows.


  • 📊 3LC- See your dataset through your model’s eyes. Capture and analyze per-sample predictions and metrics to improve your data, all in your browser*

  • 🎥 Switchlight Studio Open Beta - Revolutionize post-production lighting with AI

  • 💻 Phew AI Tab - Manage tabs with AI grouping in the sidebar

  • 🗣️ Replica Voice AI - Ethical voice AI for creators and business

  • 🚀 Simulon - Next-Gen AI and VFX Workflow for Creators

  • 📩 Superhuman Auto Email Summary - Get instant summaries of every email

*Sponsored listing


Erik Voorhees launched Venice AI, a new platform focused on privacy and user sovereignty that utilizes blockchain for payments and web3 for secure data storage.

SoftBank’s Arm is planning to launch its own AI chips in 2025 as part of CEO Masayoshi Son’s $64B investment to propel the company into a leader in the sector.

Bumble CEO Whitney Wolfe spoke on AI’s future in dating during a recent interview, envisioning an ‘AI dating concierge’ that dates other users’ AI to find ideal matches.

X unveiled a new ‘AI Audience’ feature, which allows advertisers to quickly target and better define user pools to push ads to relevant demographics.

Google DeepMind CEO Demis Hassabis said that multimodal agent systems capable of planning and achieving goals will be integrated into large AI models within the next year.

Rad AI secured $50M in Series B funding for its AI-powered tool that streamlines radiology report generation.



Get your product in front of over 600k+ AI enthusiasts

Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Get in touch today.


How would you rate today's newsletter?

Vote below to help us improve the newsletter for you.

Login or Subscribe to participate in polls.

If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.

Join the conversation

or to participate.