Good morning, AI enthusiasts. The AI world is still digesting DeepSeek’s latest R1 breakthrough, but the company is already back with another launch, taking aim at the image generation arena.

With two major releases in a matter of weeks, is tech’s latest disruptor about to upend the AI power balance?

In today’s AI rundown:

• DeepSeek launches new AI image model

• Qwen launches AI models that control devices

• Turn product photos into 3D objects

• Meta AI gets a personalization upgrade

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

DEEPSEEK

🐋 DeepSeek launches new AI image model

Image source: DeepSeek

The Rundown: Chinese AI startup DeepSeek just released Janus-Pro, a new open-source multimodal AI model that outperforms major image generation rivals like DALL-E 3 and StabIe Diffusion — coming on the heels of the company’s viral R1 launch.

The details:

• The new Janus-Pro model family generates high-quality images from text descriptions, with 1B and 7B parameter models available.

• Janus-Pro outperformed DALL-E 3 and Stable Diffusion in key industry benchmarks for image quality and accuracy, such as GenEval and DPG-Bench.

• The models were released under an MIT license, allowing developers to freely use and modify the model for commercial projects.

• The launch follows DeepSeek's R1 release, which achieved o1-level reasoning capabilities at far lower costs — shaking U.S. markets and the industry.

Why it matters: DeepSeek is the talk of the town, and the effects of R1 are being felt throughout markets as the world digests the reshaping of assumptions around development costs and capabilities. While the current panic may be an overreaction, the Chinese lab has raised questions about the U.S.'s perceived lead in the space.

TOGETHER WITH TURING

📈 Train smarter, scale faster

The Rundown: Turing’s AGI Advancement Hub provides tailored solutions to enhance your language models — helping you optimize performance, reduce costs, and align your models with real-world use cases for everything from multimodality to RLHF.

Explore how to:

• Improve reasoning, coding, and multimodal capabilities

• Deploy fine-tuned models for enterprise-grade performance

• Benchmark effectively with cutting-edge evaluation methods

Visit Turing’s AGI Advancement Hub and start optimizing today.

ALIBABA

🚀 Qwen launches AI models that control devices

Image source: Alibaba

The Rundown: Alibaba’s Qwen team just released Qwen2.5-VL, a new family of vision-language models that can interact with computers and phones, alongside new advanced capabilities in document and video analysis.

The details:

• The flagship 72B model outperforms GPT-4o and Claude 3.5 Sonnet on key benchmarks for document parsing and video understanding tasks.

• The system can analyze hour-long videos and extract specific moments while processing complex documents like invoices and forms.

• A new feature gives the AI agentic control for smartphone apps and computers, with demos including airfare booking, image editing, and code installation.

• The smaller 3B and 7B versions are freely available, with the 72B model requiring permission for large-scale commercial uses.

Why it matters: A new ‘operator’ has entered the chat — with Qwen’s computer using vision model coming just a week after OpenAI’s hyped release. Between Qwen and DeepSeek’s massive past week of releases, the gap between open and closed and China vs. U.S. models continues to feel closer than ever before.

AI TRAINING

🎨 Turn product photos into 3D objects

The Rundown: Krea AI’s real-time generation feature lets you create professional product visualizations by combining images, shapes, and AI-assisted enhancements.

Step-by-step:

1. Access Krea AI's real-time generation feature from the Generate menu.

2. Upload your product image and position it on the canvas.

3. Add shapes and painted areas where you want additional elements.

4. Fine-tune using the AI Strength slider and positioning tools.

Pro tip: Keep AI Strength low for realistic products. Increase it for creative backgrounds and effects.

PRESENTED BY INNOVATING WITH AI

💼 Launch your AI consulting career

The Rundown: Innovating with AI’s new program, AI Consultancy Project, trains AI enthusiasts with the skills and strategies to capitalize on the rapidly growing AI consulting market – which is set to 8x by 2032.

The program offers:

• The tools and frameworks to find clients and deliver top-notch services

• A 6-month plan to build a 6-figure AI consulting business

• Students getting their first AI client in as little as 3 days

Request early access to the AI Consultancy Project.

META

🧠 Meta AI gets a personalization upgrade

Image source: Meta

The Rundown: Meta just announced the rollout of new AI personalization features that let its assistant remember conversations and tap into user data across Facebook, Instagram, and WhatsApp.

The details:

• Meta AI can now remember key details from one-on-one chats, like dietary preferences and interests, to provide more tailored responses.

• The assistant will also access users' Facebook locations, Instagram viewing history, and other profile data for personalized recommendations.

• The features are launching in the U.S. and Canada across Meta's platforms with no opt-out option, though specific conversation memories can be deleted.

• ChatGPT and Gemini have also added ‘memory’ to their assistants, though limited to strictly in-chat and not Meta’s social data and integrations.

Why it matters: Meta has a wealth of social data at its disposal, and tapping into it (similar to its hyper-personalized ads) could give a unique edge with an assistant more in tune with users. However, the lack of an opt-out option feels like a major miss, especially given the company’s complicated history with user data and trust.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 Qwen2.5-1M - Alibaba’s updated models with 1M token context length

• 🦙 Llama Stack - Bring GenAI applications to market with unified APIs

• 📋 OpenAI Canvas - Now available in ChatGPT with o1 and on desktop, with new code rendering capabilities

• ⚙️ co.dev - Turn ideas into full-stack apps using natural language

💼 AI Job Opportunities

• 🤖 The Rundown - Writer (Robotics/Tech)

• 🤝 Waymo - HR Business Partner & Recruiter

• 🧠 Perplexity AI - LLM Teacher (Contract)

• 🔬 Luma AI - Senior Research Engineer

📰 Everything else in AI today

LinkedIn co-founder Reid Hoffman announced a $24.6M raise for Manas AI, a new AI-powered drug discovery platform targeting cancer treatments.

DeepSeek’s R1 surged to No.1 on Apple's App Store with 2.6M downloads, but due to reported cyberattacks, it has paused new user sign-ups outside of China.

xAI’s upcoming Grok-3 model briefly appeared to go live for some users, showcasing improved reasoning capabilities — with a full release expected this week.

Pika Labs launched v2.1 of its AI video generation model, which features advanced motion control, realistic physics, and customizable scene elements.

Apple released iOS 18.3, with the new system update turning on Apple Intelligence by default on devices — though AI summaries still remain disabled.

French government AI chatbot Lucie was suspended just days after launch following widespread errors with basic facts, despite funding from President Macron.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today at 1 PM EST and discover how your musical ideas can become part of your creative toolkit with Sara Adkins, Machine Learning Engineer and Artist in Residence at Suno.

RSVP here. Not a member yet? Join The Rundown University on a 14-day free trial.

🤝 Share The Rundown, get rewards

We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Good morning, AI enthusiasts. OpenAI’s Stargate Project be hogging the AI infrastructure headlines, but Meta and Zuck just made a Manhattan-sized move of their own.

But with Chinese models achieving similar results on shoestring budgets, are these billion-dollar bets about to reshape AI completely or reveal a massive spending bubble?

Friendly reminder: Our newsletter is still slowly moving to a new email address. Add news@daily.therundown.ai to your contacts to ensure you keep receiving our emails.

In today’s AI rundown:

• Zuckerberg announces $65B AI investment plan

• Qwen launches model upgrades, 1M token support

• How to run DeepSeek R1 offline on your computer

• Perplexity AI proposes new TikTok U.S. merger

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

META

🏗️ Zuckerberg announces $65B AI investment plan

Image source: Mark Zuckerberg on Facebook

The Rundown: Meta CEO Mark Zuckerberg just announced a massive $60-65B capital expenditure plan for 2025 focused on AI infrastructure, aiming to position Meta AI as the leading assistant and Llama 4 as the state-of-the-art model in the industry.

The details:

• The company plans to deploy roughly 1GW of compute power in 2025, building a datacenter so large it would cover a significant portion of Manhattan.

• Meta aims to amass over 1.3M GPUs by year-end, marking one of the largest AI hardware deployments globally.

• The investment represents a ~70% jump from 2024's projected spending, with Zuckerberg also predicting that Meta AI will reach 1B users this year.

• The news comes on the heels of DeepSeek R1 and OpenAI’s Stargate Project reveal, which will inject $500B into U.S. AI infrastructure projects.

Why it matters: The AI infrastructure race is on, with Meta and OpenAI both pouring unfathomable amounts of capital into the creation of giant new U.S. datacenters. The spending also comes despite DeepSeek’s recent R1 breakthrough — with the Chinese model performing on par with industry leaders despite a fraction of the training costs.

TOGETHER WITH PROJECT IDX

🚀 Boost your development with Cloud + AI

The Rundown: Project IDX offers a complete cloud development environment integrated with Google's Gemini AI — all directly within your browser.

With Project IDX, you get:

• A Linux-based virtual machine that you can access anytime, anywhere

• Native apps in browser to build and test Flutter/React Native apps with a built-in Android Emulator

• Pre-loaded environments to begin coding instantly with popular frameworks

Get up to 5 workspaces and accelerate your development today.

ALIBABA

🤖 Qwen launches model upgrades, 1M token support

Image source: Qwen

The Rundown: Alibaba's Qwen team just unveiled two new open-source models capable of processing up to 1M tokens, alongside significant speed improvements and a revamped Chat interface.

The details:

• The new Qwen2.5-1M series includes 7B and 14B parameter models, both supporting 1M token context lengths while maintaining accuracy.

• Qwen deploys a custom vLLM-inference framework, delivering up to 7x faster processing than other long-context systems.

• In tests, the Qwen-1M models outperformed other long-context models like Llama-3, GLM-4, and GPT-4 across complex long-text tasks.

• The release also includes a new Qwen Chat v0.2 upgrade, adding web search, text-to-video generation, and enhanced image capabilities.

Why it matters: Qwen’s open-source 1M models are a hint at whats to come across the industry — with Google’s Gemini (2M) and Flash 2.0 Thinking (1M) already leading the way with massive input capabilities. The context increases, paired with faster processing, open up superhuman levels of data analysis and new complex use cases.

AI TRAINING

💻 How to run DeepSeek R1 offline on your computer

The Rundown: DeepSeek recently released distilled versions of its R1 model, which can match or even outperform GPT-4o and run completely free on your computer.

Step-by-step:

1. Download LM Studio from their website for your operating system.

2. Open the app and click the magnifying glass to download a model (recommended: DeepSeek R1 Distill series).

3. Load your model from the top dropdown menu in the chat interface and start chatting.

Hot tip: These models run completely offline, so you can use them anywhere, even without internet access.

PRESENTED BY PROMPTQL

🧠 RAG with 100% accuracy

The Rundown: PromptQL transforms unpredictable AI into precise agentic intelligence, leveraging agentic planning to deliver a new paradigm of accuracy far beyond the usual 40-70% range of typical RAG and AI agents.

With PromptQL, you can build agents that:

• Use dynamic planning to build real-time strategies that adapt to changing data

• Have perfect memory to process data without context window limits

• Provide full transparency into every decision and output

Get started with PromptQL today.

PERPLEXITY

💼 Perplexity AI proposes new TikTok U.S. merger

Image source: The Rundown

The Rundown: Perplexity AI just revised its merger proposal for TikTok's U.S. operations, introducing a new structure that would give the US government a significant ownership stake in the social media giant.

The details:

• The deal would create a new company called ‘NewCo’, combining Perplexity AI and TikTok US, potentially worth as much as $300B after an IPO.

• Under the revised plan, the US government could acquire up to 50% ownership — a key sticking point in President Donald Trump’s plan for handling the sale.

• TikTok's current owner ByteDance would contribute the U.S. operations but keep the core recommendation algorithm under the proposal.

• Other rumored suitors include Elon Musk, Oracle, and Microsoft — with Trump temporarily restoring the app in the U.S. for 75 days to allow for negotiations.

Why it matters: Between its Android mobile assistant and recent API launch, Perplexity is quickly evolving from its original answer engine to an AI powerhouse — but a merger with TikTok U.S. would be by far its biggest move yet. However, with the other tech giants with deeper pockets (like Elon) circling, the plan still feels like a longshot.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 Operator - OpenAI’s first agent for in-browser tasks

• 📱 Perplexity Assistant - Android agentic assistant capable of controlling phone apps and performing complex tasks

• 📚 Citations - Allow Claude to ground answers in source documents and provide references

• 📣 WePost - Generate on-brand marketing with ease

💼 AI Job Opportunities

• 📣 The Rundown - Marketing/Media Buyer

• 🤝 Lindy AI - Head of Community

• 📊 Weights & Biases - FP&A Associate, R&D (US)

• 🧱 Databricks - Solutions Architect

📰 Everything else in AI today

ElevenLabs is reportedly raising a $250M Series C at a $3B+ valuation, with demand surging for its AI voice synthesis and dubbing technology.

Anthropic CEO Dario Amodei predicted that AI could enable humans to live 2x longer by 2030, with the tech compressing a century of research progress into 5-10 years.

xAI is reportedly developing a voice interface for its Grok iOS app with both proprietary and ElevenLabs voice options, also capable of leveraging real-time data.

OpenAI expanded Canvas functionality in ChatGPT with new rendering capabilities and o1 model support, also rolling out desktop app access across all subscription tiers.

Legendary musician Paul McCartney spoke out against the UK’s proposed AI copyright law changes, warning they could “rip off” musicians without compensation.

OpenAI CEO Sam Altman said that advances in AI will eventually require changing the social contract, with “the whole structure of society up for debate and reconfiguration.”

COMMUNITY

🎥 Join our next live workshop

Join our next workshop on Tuesday at 1 PM EST and learn how to turn ideas into original music with our next expert guest: Sara Adkins, Machine Learning Engineer and Artist in Residence at Suno.

RSVP here. Not a member yet? Join The Rundown University on a 14-day free trial.

🤝 Share The Rundown, get rewards

We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Good morning, AI enthusiasts. 2025 has already been widely declared the year of AI agents, and OpenAI just officially joined the party.

The startup’s ‘Operator’ release takes us into a new realm with mainstream AI assistants that can navigate the internet and take actions on their own — our interactions with chatbots may never be the same.

Exclusive: I got early access to test Operator, which did not disappoint. Check out my thread of demos here.

In today’s AI rundown:

• OpenAI unveils its first autonomous web agent

• Perplexity debuts new AI mobile assistant

• How to prompt o1 models better

• ‘Humanity’s Last Exam’ scales up AI benchmark

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

OPENAI

🤖 OpenAI unveils its first autonomous web agent

Image source: OpenAI

The Rundown: OpenAI just launched Operator, an AI agent that can independently navigate web browsers to complete everyday tasks — marking the company's first major step into autonomous AI assistants.

The details:

• Operator uses a new Computer-Using Agent model that combines 4o's vision capabilities with advanced reasoning to interact naturally with websites.

• OpenAI demoed the feature during a live stream, showcasing tasks like booking reservations, grocery ordering, and buying tickets to sporting events.

• OpenAI has partnered with major platforms like DoorDash, Instacart, and Uber to ensure the agent works seamlessly while respecting platform guidelines.

• Built-in safety features include user approval for purchases, automated threat detection, and "takeover mode" for sensitive info like passwords and payments.

• The research preview is currently limited to U.S. Pro users, with plans to expand to Plus, Team, and Enterprise after more safety and reliability testing.

Why it matters: While we’ve seen agentic systems popping up more frequently, OpenAI’s long-awaited move is a major step towards broadly changing the entire mindset of how we interact with AI. While there may be rough edges at first, Operator feels like the official beginning of a brand new agentic era.

TOGETHER WITH WORKOS

🔐 Protect your AI apps from bad actors

The Rundown: WorkOS Radar is a security solution that shields your AI platform from fake signups, throwaway emails, and brute force attempts — all powered by advanced device fingerprinting and real-time detection.

With WorkOS Radar, you can:

• Rapidly detect and challenge unfamiliar and suspicious devices in real time

• Stop free-tier abuse and fraudulent behavior with advanced detection

• Customize threat responses to fit your app’s exact security needs

Start integrating Radar today.

PERPLEXITY

📱 Perplexity debuts new AI mobile assistant

Image source: Perplexity

The Rundown: Perplexity just unveiled Perplexity Assistant, a free, agent-like tool for Android that can control phone apps and perform complex tasks with multimodal and voice capabilities — directly challenging voice assistants like Google’s Gemini and Siri.

The details:

• The new assistant integrates with popular apps like Uber and OpenTable to perform actions directly through voice commands or gesture controls.

• It maintains context throughout interactions, allowing users to progress from research to action — like finding restaurants and booking a table.

• The system supports multimodal interactions through voice and camera, enabling users to obtain information about their surroundings or view screen content.

• Users can replace Google's default assistant with Perplexity's solution at no cost, with the feature only available on Android for now.

Why it matters: Operator isn’t the only agent in town today, with Perplexity evolving its platform from a search/answer engine to a full-blown digital assistant. The assistant space could become a new battleground for AI firms, not just tech giants — and this Perplexity launch looks a lot like what Apple’s ‘upgraded’ Siri should actually be.

AI TRAINING

🤖 How to prompt o1 models better

The Rundown: Using delimiters and structured prompts significantly improves AI model outputs by providing clear instructions and relevant context.

Step-by-step:

1. Structure prompts using XML tags (<goal>, <context>, <format>).

2. Fill each section with relevant, specific information and define clear parameters for your desired output.

3. Test, ask follow-up questions, and optimize your results.

Pro tip: Save successful prompt structures as templates for consistent results across similar tasks. The Rundown University members can access our full workshop on effectively using ChatGPT-o1 to get the best results here.

PRESENTED BY INNOVATING WITH AI

🤝 Turn AI passion into a consulting career

The Rundown: Innovating with AI's new program, AI Consultancy Project, transforms AI enthusiasts into professional consultants — tapping into a market projected to reach $54.7B by 2032.

The 6-month program delivers:

• Proven frameworks for client acquisition and service delivery

• A step-by-step path to six-figure consulting income

• Students who land their first AI client in as little as 3 days

Click here to request early access to The AI Consultancy Project.

SCALE AI & THE CENTER FOR AI SAFETY

🧐 ‘Humanity’s Last Exam’ scales up AI benchmark

Image source: Humanity’s Last Exam

The Rundown: The Center for AI Safety and Scale AI just introduced "Humanity's Last Exam," a new AI benchmark designed to be the final frontier for testing an LLM’s academic knowledge — as current AI systems become too strong for existing tests.

The details:

• The benchmark consists of 3,000 expert-crafted questions across 100+ subjects, with contributors from over 500 institutions in 50 countries.

• Current leading AI models show surprisingly low performance on HLE, with even top systems scoring under 10% accuracy.

• Questions are in either exact-match or multiple-choice format, with 10% of the challenges incorporating multimodal analysis of text and images.

• A $500k prize pool incentivizes high-quality submissions, with top questions earning $5,000 each and co-authorship opportunities for contributors.

Why it matters: With top models routinely scoring above 90% on many of today’s key benchmarks, tests like HLE are an important way to continue scaling the ability to measure increasingly advancing AI systems. However, given the rate of progress, it likely won’t be long before we see some impressive results on these benchmarks.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 Gemini 2.0 Flash Thinking Exp - Google’s powerful new reasoning model

• ⚙️ Trae - Adaptive AI IDE that helps you ship faster

• 💻 UI-TARS - Control your computer using natural language

• 🪄 Spell by Spline - Generate full 3D scenes or worlds from a single image

💼 AI Job Opportunities

• 🎯 The Rundown - Marketing/Media Buyer

• 👋 Lindy AI - Head of Community

• 📊 Snorkel - Data Annotator (Statistics Expert Contributor)

• 🗣️ UiPath - ASR Manager

📰 Everything else in AI today

Anthropic launched Citations, a new feature in the Claude API that enables automated source attribution and verification in responses for increased accuracy.

Google’s Imagen 3.0 debuted at No. 1 in the LM Text-to-Image Arena, giving the tech giant the top spots on image and LLM leaderboards.

ByteDance is planning a $20B investment in AI infrastructure in 2025, half of which will be allocated to international data centers and partnerships with chip suppliers.

OpenAI CEO Sam Altman revealed that the upcoming o3-mini model upgrade will be available in the free tier of ChatGPT, with usage upgrades for plus users.

Hugging Face unveiled SmolVLM 256M and 500M, hailed as the world’s smallest vision language models that maintain competitive performance against larger rivals.

LinkedIn is facing a new class-action lawsuit, alleging that the company used the private messages of premium subscribers to train AI models.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today at 4 PM EST to learn about how to use DeepSeek-R1 as a powerful, cost-effective alternative for your AI projects with Dr. Alvaro Cintas, The Rundown’s AI professor.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

🤝 Share The Rundown, get rewards

We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Good morning, AI enthusiasts. In case you missed it, Cisco just announced AI Defense — a new solution that aims to become the global standard for AI security.

In a future where workforces will include AI workers — including apps, agents, robots and humanoids, understanding AI security is essential for every knowledge worker and business.

So we partnered with Cisco and Jeetu Patel, Executive Vice President and CPO at Cisco, for an exclusive Q&A on how AI Defense is tackling this challenge.

Free event: Join Cisco’s Virtual AI Summit to hear from the most influential figures in AI, including the COO of OpenAI, Scale AI, Dr. Fei-Fei Li, and more.

In today’s AI rundown:

• Cisco’s AI Defense: A new era for enterprise security

• AI Defense's two-fold data strategy for sensitive info

• Protection at scale: Network-level security integration

• The future of multi-model, multi-cloud AI security

• Why AI-forward is the only way forward

EXCLUSIVE Q&A JEETU PATEL

ANNOUNCEMENT

🛡️ Cisco’s AI Defense: A new era of AI security

Image credits: Kiki Wu / The Rundown

The Rundown: Cisco announced AI Defense, a new security solution designed to protect AI systems in a future where workforces will expand to include AI workers — including apps, agents, robots, and humanoids.

Cheung: "AI security is dominating headlines right now. But for our readers who may be new to the topic, can you explain in simple terms what AI Defense is and why this announcement is a big deal?"

Patel: "The reality is that there will be two types of companies in the future: those who are leading with AI and those that are irrelevant. Every company will be using - if not developing - thousands of AI applications, and the rapid pace of AI innovation is outperforming major security and protection concerns."

Patel added: "We developed AI Defense to protect both the development and use of AI applications. Overall, AI Defense safeguards against the misuse of AI tools, data leakage and increasingly sophisticated threats. We are taking a radical approach to address the challenges that existing security solutions are not equipped to handle."

Why it matters: Enterprises are racing to adopt AI, and due to the rapid pace of the space, traditional security solutions are falling behind. Cisco aims to make AI Defense the global standard for AI security in the increasingly AI-powered world.

STRATEGY

🧠 AI Defense's two-fold data strategy

Image credits: Kiki Wu / The Rundown

The Rundown: Cisco’s AI Defense combines two protections: monitoring third-party AI app usage and protecting sensitive data in custom AI development — addressing the growing risks of data leakage and unauthorized access.

Cheung: "A lot of enterprises are now training their own AI models on sensitive data. What specific protections does AI Defense offer that weren't possible before?"

Patel: "Enterprises are training their models on proprietary data and connecting backend data sources to create more useful AI applications. This expands the risk surface tremendously; any data connected to your AI application is now subject to leakage, poisoning, or exfiltration."

Patel added: "AI Defense takes a two-fold approach. First, it gives security teams visibility into third-party AI application usage, letting them enforce policies to limit data sharing and restrict access to high-risk solutions."

Patel added: "For enterprises developing and deploying their own AI applications, AI Defense examines user inputs and model outputs to block sensitive data leakage in real-time. This covers scenarios like an adversary attempting to extract PII or an AI application inadvertently exposing snippets of proprietary source code, for example."

Why it matters: As enterprises adopt both third-party and custom AI solutions, the risk of data leakage grows exponentially. AI Defense's two-fold approach allows organizations to innovate faster with AI while keeping their sensitive data secure. Patel goes into more depth on this strategy in his blog post.

PROTECTION

🔐 Protection at scale: How AI Defense works

Image credits: Kiki Wu / The Rundown

The Rundown: Unlike traditional security solutions, Cisco’s AI Defense is integrated directly into network infrastructure, providing consistent controls and protection at scale across all AI apps.

Cheung: "You mentioned AI Defense is 'fused into the fabric of the network' - this sounds fascinating. Can you break down what this means in practical terms and why it matters?"

Patel: "Simply put, this means that AI Defense is deeply integrated into the overall network infrastructure. By doing security at the network layer, AI Defense provides consistent controls for AI applications across the business."

Patel added: "This removes a massive burden for your AI developers. They would otherwise spend tremendous time and resources trying to secure individual applications in accordance with broader security policies."

Patel added: "This approach also leverages Cisco's existing network mesh to provide unparalleled visibility and control over AI traffic on our customers' networks. As security fuses further into the network and cloud with Hypershield, the ability of AI Defense to identify and secure east-west AI-related traffic will grow exponentially."

Why it matters: By integrating security at the network level, Cisco is changing how enterprises protect AI applications. With thousands of distributed enforcement points across clouds and servers, this approach improves security while freeing developers from application-by-application implementation.

INTO THE FUTURE

🚀 The future of AI security

Image credits: Kiki Wu / The Rundown

The Rundown: As enterprises move into a multi-model, multi-cloud world, Cisco is preparing AI Defense to tackle the growing complexity of AI security — where vulnerabilities can exist at both model and application levels.

Cheung: "We've seen AI evolve incredibly fast over the past couple years. How do you see AI Defense evolving to handle increasingly sophisticated threats, especially as AI agents become more prevalent?"

Patel: "You're right, we're already in a multi-model, multi-cloud world, and 2025 will be the year where companies start to deploy AI applications in a big way. The complexity is growing every day, and so is the risk."

Patel added: "This 'complexity' is a hard computer science problem to solve, since AI apps are built and operate differently. Ownership is fractured, meaning vulnerabilities can be at model or app level, and responsibility lives with different people – such as security practitioners, developers, users, vendors, and so on."

Why it matters: AI applications are becoming increasingly complex, and security challenges will continue to evolve. Cisco's AI Defense aims to take out the complexity by unifying security across models, applications, and stakeholders.

TRANSFORMATION

📈 Why AI-forward is the only way forward

Image credits: Kiki Wu / The Rundown

The Rundown: Jeetu Patel believes that companies who are not AI-forward will no longer be relevant in the future. With AI Defense, Cisco aims to remove the biggest barrier to adoption at the enterprise level — safety and security concerns.

Cheung: "As we enter this new era of AI adoption, how do you think AI Defense will change the way enterprises build and deploy AI applications?"

Patel: "Companies who are not AI-forward will no longer be relevant. Every application will be an AI application, and every company will be using hundreds—if not thousands—of these apps every day."

Patel added: "Our Cisco Readiness Index confirmed that one of the biggest roadblocks to AI adoption is concerns around safety and security. We built Cisco AI Defense specifically to address these concerns, and to allow enterprises to develop, deploy, and use AI applications with confidence."

Why it matters: Security concerns are the biggest barrier preventing enterprises from becoming AI-forward. Cisco's AI Defense removes this roadblock, allowing companies to confidently build and deploy more AI applications — ultimately benefiting the end users who can rely on these tools for the future of AI-enabled work.

GO DEEPER

AI SUMMIT

🤩 Free event: Join Cisco’s Virtual AI Summit

Join Cisco’s AI Summit on January 22, 2025, for an exclusive virtual broadcast featuring groundbreaking sessions with:

• OpenAI COO Brad Lightcap, to explore how AI is transforming industries like financial services, healthcare, and telecom

• Scale AI CEO Alexandr Wang on the current realities of leveraging data to build better models

• Dr. Fei-Fei Li on how AI is blurring the boundaries between virtual and physical, enabling unprecedented customization, efficiency, and engagement

• Plus more from the CEOs of Goldman Sachs, Cohere, Box, Glean, and Groq.

Register for free and witness some of the greatest minds coming together to shape the future of AI.

Good morning, AI enthusiasts. OpenAI’s AI agent era just got its unofficial start — with ChatGPT gaining the ability to schedule and manage daily tasks.

With ‘Tasks’ rolling out and mysterious 'Operator' whispers in the air, is OpenAI finally ready to move from chatbots to full-on autonomous assistants?

In today’s AI rundown:

• OpenAI launches ChatGPT ‘Tasks’

• MiniMax releases open-source, ultra-long-context LLMs

• Turn your long-form content into bite-sized clips

• Microsoft upgrades AutoGen with new multi-agent system

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

OPENAI

🤖 OpenAI launches ChatGPT ‘Tasks’

Image source: OpenAI

The Rundown: OpenAI is rolling out Tasks, a new ChatGPT beta feature that allows users to schedule reminders and recurring actions, marking the company's first step into agentic AI capabilities.

The details:

• Users can schedule one-time reminders or recurring actions, such as daily weather updates, news briefings, or periodic web searches.

• Tasks can be managed through chat or a dedicated web interface, with notifications available across desktop, mobile, and web platforms.

• ChatGPT can suggest relevant tasks based on conversation history, though users must explicitly approve any suggestions.

• Tasks will launch for Plus, Team, and Pro users in the coming days, with up to 10 active tasks at a time through a "4o with scheduled tasks" model option.

Why it matters: While reminders aren’t groundbreaking, Tasks lays the groundwork for incorporating agentic abilities into ChatGPT, which will likely gain value once integrated with other features like tool or computer use. With ‘Operator’ also rumored to be coming this month, all signs are pointing towards 2025 being the year of the AI agent.

TOGETHER WITH RUNNER H

⏰ Launch your startup in minutes

The Rundown: 2025 is the year of AI Agents— and Runner H's state-of-the-art agent platform (backed by a historic $220M pre-seed round) allows you to launch your own business in just a few clicks.

With Runner H, you can:

• Launch your business with automated setup and workflows

• Leverage AI agents for competitor research, pricing, website building and more

• Transform text instructions into real-world business operations

Watch this video to learn more and join the waitlist to be among the first to test an agent that transforms the way you work.

MINIMAX

⚡️ MiniMax releases open-source, ultra-long-context LLMs

Image source: Minimax

The Rundown: Chinese AI lab Minimax just launched two new open-source AI models that leverage a new ‘Lightning Attention’ approach that allows for massive context windows of 4M tokens while maintaining speed and performance.

The details:

• The release includes a 456B parameter base language model (MiniMax-Text-01) and a multimodal model (MiniMax-VL-01).

• Both models can process sequences up to 4M tokens, dramatically exceeding current industry standards of 128K-256K tokens.

• The models perform comparable to top models on academic benchmarks, outperforming all open-source models on long-context tasks.

• The company also offers API access at notably low rates, with input tokens at $0.2/million and output tokens at $1.1/million.

Why it matters: As AI development shifts toward autonomous agents with extensive memory and context processing needs, MiniMax's ultra-long context could be revolutionary. Open-sourcing these models, combined with competitive API pricing, could kickstart an aggressive innovation push in the AI agent ecosystem.

AI TRAINING

✂️ Turn your long-form content into bite-sized clips

The Rundown: HeyGen’s Instant Highlights feature transforms long videos into engaging highlight reels and shareable clips.

Step-by-step:

1. Access the Labs section from HeyGen and find Instant Highlights.

2. Add your video through URL or direct upload (supports up to 10GB).

3. Customize the highlight settings by selecting your preferred clip duration, aspect ratio, and if you want captions.

4. Download or share your favorite clips directly.

Pro tip: Use the "Share highlight list" feature to collaborate with your team on selecting the best clips.

PRESENTED BY BELVA AI

💻 Turn developers into architects with AI

The Rundown: Belva AI's Engineer is an AI-powered development environment that automates coding workflows while keeping humans in control — combining automated code generation, UML transformation, and intelligent debugging in one platform.

Leverage Engineer in your development process to take advantage of:

• Automated code generation & hosting

• AI-generated diagram-driven development

• Limitless context capabilities for LLMs

Get started with Engineer and simplify your workflow today.

MICROSOFT

🤖 Microsoft upgrades AutoGen with new multi-agent system

Image source: Microsoft

The Rundown: Microsoft just released AutoGen v0.4, the latest update to its open-source agentic framework — alongside the introduction of Magnetic-One, a new orchestration system that coordinates multiple AI agent specialists to handle complex tasks.

The details:

• Magnetic-One introduces four agents: WebSurfer for web navigation, FileSurfer for local file management, and Coder and ComputerTerminal for coding.

• V0.4’s event-driven messaging system enables async communication between agents, allowing for more flexible, customizable, and complex workflows.

• The release also includes AutoGen Studio for low-code development, AutoGen Bench for performance testing, and upgraded monitoring tools.

• The system also remains LLM-agnostic, working with different language models while defaulting to GPT-4o integration.

Why it matters: While the world has just started dipping its toes into the AI agent boom, the next steps are already being taken to enable multi-agent systems that open the door to tackling complex applications and tasks. Our own personal agentic teams are right around the corner — and human workflows will never be the same again.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 GitHub Copilot - AI coding assistant, now available without a waitlist

• 🤝 Reachy AI - AI agent to automate outreach on LinkedIn

• 🌝 Fullmoon - Chat with private and local large language models

• 🔵 Shapen - Create 3D models from images with AI

💼 AI Job Opportunities

• 💻 EliseAI - Senior Software Engineer, Healthcare*

• 🦺 xAI - HSE Coordinator

• 🤝 OpenAI - Engagement Manager

• 🚀 Palantir Technologies - Forward Deployed Software Engineer

* Sponsored listing

📰 Everything else in AI today

President Joe Biden signed a new executive order allowing AI companies to build data centers on DoD and Energy sites to strengthen domestic AI infrastructure development.

Amazon is reportedly working to transform its Alexa assistant into an AI-powered agent, but it has faced issues like hallucinations and slow response speeds that, if launched, risk damaging the company’s brand and reputation.

The Mayo Clinic partnered with Microsoft and Cerebras to develop AI foundational models for medical image and genomic data analysis, hoping to revolutionize personalized medicine with faster diagnostics and precise treatment predictions.

OpenAI added GIP chairman and Blackrock executive Adebayo Ogunlesi to the company’s Board of Directors, bringing experience in infrastructure investment and global finance expertise.

French AI startup Bioptimus secured $41M in funding to develop a “GPT for biology” foundation model to simulate biological systems and predict disease outcomes.

Microsoft announced a new multiyear partnership with education giant Pearson to transform AI learning and workforce development, with plans to deliver AI-powered skilling solutions and certifications globally.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today at 2 PM EST to learn how to create custom styles with Sofia Lopez Martin, Freepik's expert host.

RSVP here. Not a member yet? Join The Rundown University on a 14-day free trial.

🤝 Share The Rundown, get rewards

We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.

Your current referral count: {{rp_num_referrals}}
What you need: {{rp_num_referrals_until_next_milestone}} more referrals until receiving {{rp_next_milestone_name}}

👀 Get featured in the world’s largest AI newsletter

Interested in sponsoring us and reaching {{active_subscriber_count}} AI enthusiasts, tech executives, investors, engineers, and business owners around the world? Get in touch today.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Welcome, AI enthusiasts.

While tech giants continue to pour billions into AI development, a Chinese startup just proved you can build an open-source powerhouse on a tiny budget.

DeepSeek’s V3 model competes with the top models at a fraction of the cost — and the walls between open and closed AI are crumbling faster than ever. Let’s get into it…

In today’s AI rundown:

• DeepSeek-V3 rewrites open-source AI playbook

• OpenAI reveals restructuring plans for next AI phase

• Turn your images into any style

• Stanford AI brings natural gestures to digital avatars

• 5 new AI tools & 5 new AI jobs

• More AI & tech news

Read time: 4 minutes

Friendly reminder: We’re moving to a new email sending address. Make sure to add news@daily.therundown.ai to your contacts to ensure you keep receiving our emails.

LATEST DEVELOPMENTS

DEEPSEEK

🤖 DeepSeek-V3 rewrites open-source AI playbook

Image source: DeepSeek

The Rundown: Chinese AI startup DeepSeek just released DeepSeek-V3, a new powerhouse language model that sets new benchmarks in the open-source AI world with performance rivaling industry giants at a fraction of the cost.

The details:

• V3 uses a Mixture-of-Experts architecture and maintains speed and cost-effectiveness despite its massive 671B parameter size.

• The training was completed in just two months at an estimated $5.57M, dramatically less than the reported $500M+ spent on models like LLaMA 3.1.

• The model shows exceptional strength in math and Chinese language tasks while matching or exceeding closed models across most benchmarks.

• V3 has been critiqued for identifying as ChatGPT in conversations, which may be due to significant GPT-generated content used in its training dataset.

Why it matters: The gap between open and closed AI models has never been smaller. Chinese models continue to prove that the U.S. chip restrictions are failing to slow progress, and V3’s benchmarks show that open-source, high-performance models are achievable without the massive resources of other tech giants.

TOGETHER WITH VANTA

⚒️ Building startups in the age of AI

The Rundown: Join Vanta for a fireside chat with CEO Christina Cacioppo and "The Lean Startup" author Eric Ries to gain insights on navigating product-market fit, delegation, and scaling startups in the age of AI.

The Jan. 14 session will cover:

• Key takeaways from Christina and Eric’s entrepreneurial journeys

• How the principles of “founder mode” and lean methodology intersect

• The ways AI is transforming early-stage growth and startup success

Secure your spot today.

OPENAI

🔄 OpenAI reveals restructuring plans for next AI phase

Image source: Getty Images

The Rundown: OpenAI officially published its plans to transform into a public benefit corporation, posting a blog detailing the company’s controversial shift from its nonprofit structure to enable massive funding abilities for future AI development.

The details:

• The restructuring will convert OpenAI's for-profit arm into a Delaware-based PBC, with the original nonprofit gaining ‘significant’ shares in the new entity.

• OpenAI said the result would be ‘one of the best-resourced non-profits in history,’ enabling the pursuit of charitable goals in health care, education, and science.

• The process follows the startup’s $6.6B funding round at a $157B valuation, which reportedly hinged on the restructuring.

• Elon Musk sued OpenAI in December to prevent the move, with California nonprofit Encode also pushing for a pause in a new filing.

Why it matters: OpenAI’s convoluted structure has been criticized since the Nov. 2023 board drama, but as Elon and co. continue to hold the startup’s feet to the fire, the restructure process seems imminent. Threading the needle between a mission-driven focus and a for-profit endeavor is about to become a delicate balancing act.

AI TRAINING

🎨 Turn your images into any style

The Rundown: Google’s new Whisk tool allows you to create custom-stylized versions of any photo by combining subjects, scenes, and artistic styles without writing prompts.

Step-by-step:

1. Visit Google Labs and open the Whisk experiment (free to use).

2. Click "Start from scratch," upload your subject and scene, and choose a style.

3. Let AI generate multiple consistent variations of your image.

4. Refine results by adding specific details through the text box.

Pro tip: Use the "+" button to add multiple subjects or scenes for more complex compositions.

PRESENTED BY INNOVATING WITH AI

🤝 Turn AI passion into a consulting career

The Rundown: Innovating with AI's new program, AI Consultancy Project, transforms AI enthusiasts into professional consultants — tapping into a market projected to reach $54.7B by 2032.

The 6-month program delivers:

• Proven frameworks for client acquisition and service delivery

• A step-by-step path to six-figure consulting income

• Students who land their first AI client in as little as 3 days

Click here to request early access to The AI Consultancy Project.

AI RESEARCH

🎭 Stanford AI brings natural gestures to digital avatars

Image source: Stanford University

The Rundown: Stanford researchers just developed an AI model that enables digital avatars to produce natural, context-appropriate gestures while speaking — marking advances in making virtual characters more lifelike and expressive.

The details:

• Training data included 1,000 hours of audiobooks and 60 hours of motion capture to understand relationships between speech, text, and body language.

• The system can generate appropriate gestures from text or audio input while detecting emotions from motion patterns alone.

• Using a new pre-training approach, the AI requires less training data than past models while achieving better performance on gesture generation benchmarks.

• The model shows particular strength in matching gestures to speech patterns, such as emphasizing specific words through corresponding body movements.

Why it matters: Zoom CEO Eric Yuan has said we will soon send avatars to meetings in our place. Having realistic motion and gestures is a crucial step towards crafting truly reality-bending characters. Digital avatars are about to become ubiquitous across industries, and these advances could open up a whole new world of use cases.

NEW TOOLS & JOBS

Trending AI Tools

• ⚙️ ReactAI - Create fully-functional React components in seconds with no API key required

• 🔎 Hika - AI-powered knowledge search tool that transforms exploration and simplifies complex topics into interactive sections

• 📚 Scourhead - Free, open-source AI agent that scours the web, organizes data, and delivers results in a spreadsheet

• 🤖 SEObot - Produce useful, non-spammy AI-generated blog content with auto-linking, keyword research, image generation, and fact-checking

• 📊 Shoonya - Specialized foundation models fine-tuned for commerce use cases

New AI Job Opportunities

• 🤖 Meta - Software Engineering Manager - FAIR

• 🛡️ Shield AI - Product Marketing Manager Consultant

• 🚗 Waymo - Senior Program Manager, Commercial Operations

• 🔬 Scale AI - SEAL Research Scientist, Scalable Oversight

• 🏙️ Metropolis - Principal Software Engineer

QUICK HITS

‘Godfather of AI’ Geoffrey Hinton raised his odds of human extinction risk due to superintelligent AI development, revising his estimate to 20% within three decades while calling for stronger government regulation of the technology.

OpenAI and Microsoft reportedly agreed on a concrete metric for achieving artificial general intelligence, with a newly revealed document from 2023 defining AGI as an AI system capable of generating $100B in annual profits.

Meta unveiled a vision for AI-generated characters to become active social media users across its platforms, with plans to expand capabilities from profile creation to content generation and live interactions.

Chinese robotics firm Unitree demoed B2-W, a rideable robot dog capable of carrying humans across challenging terrain in addition to acrobatic maneuvering and stability control.

Toyota’s AI-powered humanoid robot CUE6 set a new Guinness World Record for the longest basketball shot by a robot, sinking an 80-footer on its second try.

THAT’S A WRAP

See you soon,

Rowan, Joey, Zach, and Alvaro—aka The Rundown Team

Welcome, AI enthusiasts.

Today is our last email before we take our team-wide week-long holiday break (our first break in nearly two years — crazy!).

So in the spirit of rest and relaxation, we've got a jam-packed rundown for you today with big news from Google DeepMind, Genesis, Apptronik, and more. Let's get into it…

In today’s AI rundown:

• Google releases experimental 'reasoning' AI

• The first generative AI physics simulator

• Turn your articles into video podcasts

• Google partners with Apptronik on humanoid robots

• 6 new AI tools & 5 new AI jobs

• More AI & tech news

Read time: 4 minutes

Friendly reminder: We’re moving to a new email sending address. Make sure to add news@daily.therundown.ai to your contacts to ensure you keep receiving our emails.

LATEST DEVELOPMENTS

GOOGLE DEEPMIND

 ✨ Google releases experimental 'reasoning' AI

Image source: Google

The Rundown: Google just released Gemini 2.0 Flash Thinking Experimental, a new AI model that pauses to "think" through complex problems like OpenAI's o1 model, but is free-to-use and works faster.

The details:

• The model explicitly shows its thought process while solving problems, similar to other reasoning models like OpenAI's o1.

• Built on Gemini 2.0 Flash, early users report significantly faster performance than competing reasoning models.

• The model increases computation time to improve reasoning, leading to longer but potentially more accurate responses.

• The model is now ranked #1 on the Chatbot Arena across all categories and is freely available through AI Studio, the Gemini API, and Vertex AI.

Why it matters: The race for better AI reasoning capabilities is intensifying, with Google joining OpenAI and others in exploring new approaches beyond just scaling up model size. While OpenAI continues to increase pricing for their top-tier models, Google continues taking the opposite approach by making its best AI freely accessible.

TOGETHER WITH NORTHERN DATA GROUP

🆕 Be the first to access NVIDIA's next-gen chip

The Rundown: Northern Data is the first European GenAI Cloud Provider to offer instant access to NVIDIA H200 GPUs — providing a game-changing advantage for your AI development and deployment.

Benefits include:

• Immediate access to revolutionary AI hardware

• The ability to supercharge AI workloads with cutting-edge performance

• A new competitive technical edge to stay ahead of the competition

Pre-register today to reap the rewards of the H200’s top-tier AI compute power.

GENESIS

🌎 The first generative AI physics simulator

Image source: Genesis

The Rundown: A team of researchers across 20 separate labs just unveiled Genesis, an open-source physics engine that combines generative AI with ultra-fast simulations, potentially transforming how AI learns to interact with the physical world.

The details:

• Genesis runs 430,000 times faster than real-time physics, achieving 43 million FPS on a single RTX 4090 GPU.

• It’s built in pure Python, it's 10-80x faster than existing solutions like Isaac Gym and MJX.

• The platform can train real-world transferable robot locomotion policies in just 26 seconds.

• The platform is fully open-source and will soon include a generative framework for creating 4D environments.

Why it matters: By enabling AI to run millions of simulations at unprecedented speeds, Genesis could massively accelerate robots' ability to understand our physical world. Open-sourcing this tech, along with its ability to generate complex environments from simple prompts, could spark a whole new wave of innovation in physical AI.

AI TRAINING

🎥 Turn your articles into video podcasts

The Rundown: HeyGen’s new ‘Podcast‘ feature allows users to turn an article, research paper, or document into an engaging two-person video discussion with AI avatars (similar to NotebookLM but with real faces).

Step-by-step:

1. Visit HeyGen's website and select "Video Podcast" from the dashboard.

2. Upload your content (PDF up to 20MB or website URL).

3. Pick two AI presenters and customize video settings (duration, aspect ratio, captions).

4. Click submit and get your video in minutes.

Pro tip: Well-structured content leads to more natural-sounding discussions. You can use an AI tool like ChatGPT to structure content into a better script beforehand.

GOOGLE & APPTRONIK

🤖 Google partners with Apptronik on humanoid robots

Image source: Apptronik

The Rundown: Google DeepMind just announced a strategic partnership with Apptronik, an Austin-based humanoid robotics company, to combine advanced AI with cutting-edge hardware in the pursuit of versatile humanoid robots.

The details:

• Apptronik brings nearly a decade of robotics expertise, including the development of NASA's Valkyrie Robot and their current humanoid, Apollo.

• Apollo stands 5'8", weighs 160 pounds, and is designed for industrial tasks while safely working alongside humans.

• The partnership will leverage Google DeepMind's AI expertise, including their Gemini models, to enhance robot capabilities in real-world environments.

• This marks Google's return to humanoid robotics after selling Boston Dynamics to SoftBank in 2017.

Why it matters: Seven years after selling Boston Dynamics, Google is re-entering humanoid robotics — this time through AI rather than hardware. This partnership could give DeepMind's advanced AI models (like Gemini) a physical form, potentially bringing us closer to practical humanoid robots that can work alongside humans.

NEW TOOLS & JOBS

Trending AI Tools

• 🎥 Kling AI v1.6 - An update to the popular AI video generator, which includes improved prompt adherance, professional modes, and more

• ✨ Gemini 2.0 Flash Thinking - Google DeepMind’s latest free-to-try reasoning model that competes with ChatGPT o1

• 🧊 Backflip AI - Turns text into 3D AI-generated designs

• ✏️ tldraw computer - An infinite canvas for natural language computing

• 💻 ModernBERT - A family of SOTA encoder-only models with major improvements over older generation encoders

• 👀 Microsoft Copilot Vision - An AI companion that can see what you do in the Edge browser (now rolling out to U.S. Copilot Pro subscribers on Windows)

New AI Job Opportunities

• 🛠️ OpenAI - Solutions Engineer, Pursuits

• 📋 Palantir Technologies - Talent Coordinator

• 📊 Weights & Biases - Senior Product Manager, Models

• 🔒 Mistral AI - Software Engineer, Security

• 🎯 Meta - Product Marketing Manager Intern, MBA

QUICK HITS

OpenAI's Alec Radford, lead author of GPT and other key technologies, announced his departure to pursue independent research following the recent exits of several senior research leaders from the company.

Anthropic published new best practices for building AI agents, emphasizing simple, composable patterns over complex frameworks and sharing insights from customer implementations across industries.

Meta hinted at speech capabilities and advanced reasoning in Llama 4 in a recent blog post, alongside plans for business-focused AI agents for customer support and commerce in 2025.

Perplexity acquired Carbon, a data connectivity startup, to help users connect apps like Notion and Google Docs directly to its AI search platform.

Microsoft AI announced that Copilot Vision — the company’s AI companion that can see your browser and interact with you in real-time — is rolling out to U.S. Copilot Pro users on Windows.

OpenAI expanded ChatGPT's app integration to support more coding platforms, including JetBrains IDEs, terminals, and productivity apps like Apple Notes and Notion.

Anthropic published new research showing their AI models can engage in "alignment faking" — appearing to comply with new training while preserving original preferences.

THAT’S A WRAP

See you in a week, and have a happy holiday! 🎄

Rowan, Joey, Zach, and Alvaro—aka The Rundown Team

Welcome, AI enthusiasts.

OpenAI just turned back the clock with a surprisingly retro way to access ChatGPT — and your parents’ landline will work just as well as your smartphone.

With a new 800 ‘hotline’ number and WhatsApp integration, OpenAI just made accessing advanced AI as simple as dialing a friend. Let’s get into it…

In today’s AI rundown:

• ChatGPT gets a new phone number

• GitHub Copilot goes freemium

• Create infinite looping AI videos

• Perplexity hits $9B valuation in mega-round

• 5 new AI tools & 5 new AI jobs

• More AI & tech news

Read time: 4 minutes

LATEST DEVELOPMENTS

OPENAI

☎️ ChatGPT gets a new phone number

Image source: OpenAI on YouTube

The Rundown: OpenAI just launched a surprising new way to access ChatGPT — through an old-school 1-800 number and also rolled out a new WhatsApp integration for global users during Day 10 of the company’s livestream event.

The details:

• US users can now dial 1-800-CHATGPT to have voice conversations with the AI assistant, and they will receive 15 minutes of free calling time per month.

• The phone service works on any device, from smartphones to vintage rotary phones — allowing accessibility without requiring modern tech.

• A parallel WhatsApp integration also lets international users text with ChatGPT, though with feature limitations compared to the main app.

• The WhatsApp version runs on a lighter model with daily usage caps, offering potential future upgrades like image analysis.

Why it matters: The days of operators or hotlines are long gone, but OpenAI’s latest move brings that style of information back to the AI age in an accessible way. By meeting users where they are — whether on WhatsApp or a landline phone — OpenAI is expanding AI’s reach beyond the tech-savvy crowd.

TOGETHER WITH READPARTNER

🕒 Turn hours of reading into minutes

The Rundown: ReadPartner combines news automation and AI-powered summarization to help you master your inbox, conquer content, and reclaim lost time — boasting a proven history of industry leading performance.

ReadPartner will help you:

• Summarize content from emails, videos, documents, and more with one click

• Automate consumption with personalized news digests straight to your inbox

• Save 28+ hours of valuable time every month

Sign up now with a New Year’s special to start transforming your browsing experience.

MICROSOFT & GITHUB

🆓 GitHub Copilot goes freemium

Image source: GitHub

The Rundown: Microsoft-owned GitHub just announced a free tier of its AI coding Copilot available in its VS Code editor, marking a major shift in AI coding accessibility as the company celebrates a milestone of 150M developers on the platform.

The details:

• The new free tier offers 2,000 monthly code completions and 50 chat messages, integrated directly into VS Code and GitHub's dashboard.

• Users can access Anthropic's Claude 3.5 Sonnet or OpenAI's GPT-4o models, with premium models (o1, Gemini 1.5 Pro) remaining exclusive to paid tiers.

• Free features include multi-file editing, terminal assistance, and project-wide context awareness for AI suggestions.

• GitHub also announced its 150M developer milestone, up from 100M in early 2023.

Why it matters: GitHub has lofty ambitions to reach 1B developers globally, and removing price barriers would go a long way toward onboarding the masses and preventing existing users from flocking to the other free options on the market. The future of AI coding is increasingly looking more like a fundamental free utility than a premium tool.

AI TRAINING

🎬 Create infinite looping AI videos

The Rundown: Sora’s AI video generator lets you transform short videos into a seamless loop that plays continuously without noticeable cuts or transitions.

Step-by-step:

1. Log into Sora with your Plus/Pro subscription.

2. Generate your base video by describing continuous motion (e.g., "dogs running across the screen").

3. Click the Loop button and choose your loop type (Short/Normal/Long).

4. Export your seamless loop for websites, social media, or digital displays.

Pro tip: For best results, choose subjects with natural continuous motion and avoid sudden changes in direction or speed.

PRESENTED BY INNOVATING WITH AI

🤝 Turn AI passion into a consulting career

The Rundown: Innovating with AI's new program, AI Consultancy Project, transforms AI enthusiasts into professional consultants — tapping into a market projected to reach $54.7B by 2032.

The 6-month program delivers:

• Proven frameworks for client acquisition and service delivery

• A step-by-step path to six-figure consulting income

• Students who land their first AI client in as little as 3 days

Click here to request early access to The AI Consultancy Project.

PERPLEXITY

💰 Perplexity hits $9B valuation in mega-round

Image source: Bloomberg

The Rundown: AI search startup Perplexity just secured a massive $500M funding round, which triples its valuation to $9B, as the company continues to expand its challenge to traditional search engines.

The details:

• The company's valuation has skyrocketed from $1B in April to $9B in this latest round, and the rise has come despite lawsuits from major publishers.

• Since its launch in 2022, Perplexity has attracted over 15M active users, with recent feature additions including one-click shopping and financial analysis.

• The startup has inked revenue-sharing deals with major publishers like Time and Fortune to address content usage concerns.

• Perplexity also acquired Carbon, a data connectivity startup, to enable direct integration with platforms like Notion and Google Docs.

Why it matters: Perplexity’s rise is evidence of the shakeup underway in the search landscape, with other leaders like ChatGPT also rushing to turn their own platforms into modern search engines. With heavyweights like Google also revamping search for the AI age, will Perplexity be able to stay a step ahead of the industry giants?

NEW TOOLS & JOBS

Trending AI Tools

• 🎥 Meta Video Seal - Open-source tool embedding invisible, durable watermarks in videos to ensure authenticity

• 🏋️ Impakt AI App - AI coach that can see, talk, and instruct your workouts to guide you toward your personal fitness goals

• 👨‍💻 Tempo Labs - AI-powered visual editor for React, giving PMs, designers, and engineers the ability to collaborate visually on code

• 🎵 TemPolor - Royalty-free, AI-powered music platform designed to empower content creators with customizable music that enhances storytelling

• 🦦 Otterly AI - Monitor brand and link visibility on ChatGPT, Perplexity, and AIO

New AI Job Opportunities

• 📊 Databricks - Director TPM, Infrastructure

• 🌍 Anyscale - Regional Sales Director

• 💼 Runway - Sr. Technical Recruiter

• 🖥️ Abridge - IT Support Manager

• 📱 Siena - Social Media & Community Manager

QUICK HITS

Free workshop by Section: Get proficient in prompting, learn solid AI use cases, and get confident using AI in just one hour. Sign up here.*

Microsoft reportedly acquired nearly 500,000 Hopper GPUs from Nvidia in 2024, emerging as the chipmaker’s biggest customer and nearly doubling the purchases from rivals like Meta and ByteDance.

Magnific AI released Magic Real, a new image generation model focusing on realistic outputs for professionals in architecture, photography, film, and interior design.

Odyssey revealed Explorer, a new generative world model that converts images into 3D environments, also adding Pixar co-founder Ed Catmull to its Board of Directors.

Open Vision Engineering introduced Pocket, a $79 physical AI-powered voice recorder that captures, transcribes, and organizes conversations.

Runway launched a new talent network platform, connecting AI filmmakers and production houses with brands and studios seeking expertise in the sector.

The U.S. Department of Homeland Security unveiled DHSChat, an internally developed AI chatbot deployed on secure infrastructure for its 19,000 employees and agency users.

*Sponsored listing

THAT’S A WRAP

SPONSOR US

Get your product in front of over 800k+ AI enthusiasts

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

See you soon,

Rowan, Joey, Zach, and Alvaro—aka The Rundown Team