Good morning, AI enthusiasts. Sam Altman just dropped OpenAI’s new model roadmap and scrapped the standalone release of o3 — but something much bigger is taking its place.

With a new plan to unify the company’s tools and models under an upcoming GPT-5 umbrella, could the days of the fragmented AI ecosystem finally end?

Reminder: Our next workshop is today at 4 PM EST, where you’ll learn how to build your own AI-powered personal assistant to streamline your daily tasks! RSVP here.

In today’s AI rundown:

• Sam Altman details GPT-5 roadmap, scraps o3 launch

• Adobe launches IP-safe AI video generator

• Repurpose video transcripts across platforms

• OpenAI expands Model Spec with focus on intellectual freedom

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

OPENAI

🗺️ Sam Altman details GPT-5 roadmap, scraps o3 launch

Image source: Sam Altman (@sama) on X

The Rundown: CEO Sam Altman just posted a roadmap update for OpenAI’s upcoming models, canceling its anticipated o3 release in favor of an integrated GPT-5 system that will unify and simplify the companies’ AI offerings.

The details:

• GPT-5 will incorporate o3's capabilities alongside other OpenAI tech, with a tiered system offering different intelligence levels for free, Plus, and Pro users.

• Before GPT-5, the company will release GPT-4.5 (codenamed "Orion") as its final non-chain-of-thought model in the coming weeks.

• Free users will have unlimited access to GPT-5 at "standard intelligence," while paid tiers unlock higher performance and advanced tools.

• Altman provided an ETA of ‘weeks’ for 4.5 and ‘months’ for 5, with o3 no longer being released as a standalone model.

Why it matters: The model landscape (and naming conventions) have become increasingly difficult to follow, making for a poor user experience. This new unified structure sounds like GPT-5 will automatically select the best models and tools for a specific task, potentially reshaping how intelligence is packaged and delivered to users.

TOGETHER WITH UNSTRUCTURED

🥋 Master your unstructured GenAI data

The Rundown: We’re all building our GenAI stacks on the fly. Join Unstructured & Elastic for a live session to accelerate your learning curve and learn the latest techniques for fueling your apps and pipelines with all your document data.

Join the live webinar on Jan. 22 to learn how to:

• Build streamlined ETL pipelines for your unstructured GenAI data

• Balance good/fast/affordable parsing for your most complex documents

• Parse and search everything once your data is within Elastic

• Unlock internal business data with a live code demo of actionable steps

Click here to register and receive the recording, even if you can’t attend live.

ADOBE

🎨 Adobe launches IP-safe AI video generator

Image source: Adobe

The Rundown: Adobe just released its Firefly Video Model in public beta, positioning it as the industry's ‘first commercially safe AI video generation tool’ — along with new pricing tiers starting at $9.99 for its expanded Firefly creative application suite.

The details:

• The new system can generate 1080p video clips from text or images, with precise camera control, shot framing, and motion graphics capabilities.

• The model is trained on licensed Adobe Stock and public domain content, and the company emphasizes that it will never be trained on customer generations.

• Adobe is launching two new subscription tiers: Standard ($9.99/month for 20 videos) and Pro ($29.99/month for 70 videos).

• Other upgrades include Translate and Lip Sync for audio, Scene to Image for 3D structure references, and broader integrations with other Adobe platforms.

Why it matters: There is no shortage of competition for the AI creative suite, but few companies have focused directly on being commercially safe — a differentiator that could make Firefly the go-to choice for brands and agencies that need legally safe AI-generated content at scale.

AI TRAINING

📝 Repurpose video transcripts across platforms

The Rundown: Claude can transform your video transcripts into multiple content pieces while maintaining your unique voice and style.

Step-by-step:

1. Gather your transcript and 2-3 samples of your best content.

2. Train Claude with your detailed style-matching prompt by instructing it to analyze your examples.

3. Request to transform your transcript into focused, voice-matched content for one or multiple social media platforms.

Pro tip: Save time by creating a custom Claude Project trained specifically on your voice and style. The Rundown University member can access the top prompts here.

PRESENTED BY INNOVATING WITH AI

💼 Start your career as an AI Consultant

The Rundown: Innovating with AI’s new program, AI Consultancy Project, equips AI enthusiasts with all the resources to capitalize on the rapidly growing AI consulting market – which is set to 8x to $54.7B by 2032.

The program offers:

• Tools and framework to find clients and deliver top-notch services

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

• Student landing their first AI client in as little as 3 days

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

OPENAI

📋 OpenAI expands Model Spec with focus on intellectual freedom

Image source: OpenAI

The Rundown: OpenAI just released an expanded Model Spec that outlines how models should handle controversial topics, emphasizing truth, user customization and intellectual freedom while maintaining key safety guardrails.

The details:

• The 63-page specification introduces a "chain of command" where platform rules precede developer and user preferences.

• After feedback requesting a "grown-up mode," OpenAI is exploring ways to allow types of adult content while maintaining strict bans on harmful material.

• The company is combatting ‘AI sycophancy’ by training models to give honest feedback instead of empty praise and avoiding agenda-seeking responses.

• The Model Spec is released under a CC0 license, allowing other AI companies to adopt and modify these guidelines for their own systems.

Why it matters: Another common critique of OpenAI is the frequent restrictions around controversial topics or conversations. With open source and unfiltered models like Grok gaining in popularity, it appears ChatGPT is going down the less restrictive, ‘truth-seeking’ route — a move even Elon Musk should ideologically approve of.

QUICK HITS

🛠️ Trending AI Tools

• ⚡Deepgram - Introducing Nova-3, the first voice AI model with real-time multilingual speech-to-text*

• 🧠 Perplexity Sonar - New upgraded model with 10x faster response speed

• 🧮 DeepScaleR - Open-source 1.5B model with exceptional math reasoning

• 🧊 3D AI Studio V3 - Generate, organize, and edit 3D models
  

  *Sponsored Listing

💼 AI Job Opportunities

• 🎯 The Rundown - Marketing/Media Buyer

• 📦 Coreweave - Warehouse Associate

• 🛠️ Shield AI - Staff Engineer, DevOps

• 💼 Databricks - Startup Account Executive

📰 Everything else in AI today

Apple analyst Ming-Chi Kuo revealed that Apple is exploring humanoid and non-humanoid robots for its smart home ecosystem, though mass production isn't expected before 2028.

Sam Altman said that OpenAI is planning to extend access to its Deep Research tool to all ChatGPT tiers, with 2 uses per month for free users and 10 for Plus users to start.

Thomson Reuters secured a landmark AI copyright legal victory, with a judge ruling that Ross Intelligence’s use of copyrighted content for AI training constituted infringement.

Midjourney founder David Holz teased that the company has two hardware projects currently in development, with ‘one that goes on you’ and ‘one that you go inside of.’

AI infrastructure startup fal secured a $49M Series B to expand its video-focused generative media platform, which already processes over 100M daily inference requests for enterprise customers including Quora and Canva.

Glean launched Glean Agents, a new platform allowing enterprises to build and deploy custom assistants with access to company and internet data.

COMMUNITY

🎥 Join our next live workshop

Join our workshop today at 4 PM EST to learn how to build your own AI-powered personal assistant to streamline your daily tasks with Dr. Alvaro Cintas, The Rundown’s AI professor.

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. While tech giants race to build more powerful AI, world leaders clash over who gets to write the rules.

The AI Action Summit just wrapped up in Paris, exposing some deep new fractures between allies — with billions in funding and global dominance at stake.

In today’s AI rundown:

• Global tensions rise at Paris AI summit

• Perplexity drops blazing new Sonar model

• Master pre-market analysis with AI insights

• YouTube supercharges creators with AI expansion

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

AI ACTION SUMMIT

🌍 Global tensions rise at Paris AI summit

Image source: BBC

The Rundown: The AI Action Summit in Paris concluded on Tuesday, raising global tensions as the U.S. and UK refused to sign a multinational declaration and European leaders announced ambitious investment plans to compete with American dominance.

The details:

• U.S. Vice President J.D. Vance warned against AI overregulation, saying that the U.S. would dominate AI development by controlling chips, software, and rules.

• The UK and the U.S. declined to sign a declaration for open, ethical AI, citing national security concerns and disagreements over governance.

• Anthropic CEO Dario Amodei called the summit a ‘missed opportunity,’ highlighting concerns over accelerating AI progress and security risks.

• EC President von der Leyen announced a €200B AI investment initiative, positioning Europe as an open-source alternative to U.S. AI development.

Why it matters: The AI summit revealed a widening rift in approaches to AI governance. With the U.S. and the previously safety-focused UK not committing to the summit’s pledge and China now part of the group of signees, AI is becoming a massive global policy issue — with the ability to reshape power balances and alliances quickly.

TOGETHER WITH H COMPANY

🤖 What’s new in the world of AI agents?

The Rundown: H Company’s Runner H, backed by a historic $220M pre-seed investment, is your new go-to AI agent for automating tasks that usually take hours in just a few clicks.

With Runner H, you can:

• Automate complex workflows without writing any code

• Create web automations seamlessly

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

• Transform text instructions into real business operations

Watch this video and follow H Studio on X to be among the first to test an AI agent that transforms the way you work.

PERPLEXITY

💨 Perplexity drops blazing new Sonar model

Image source: Perplexity

The Rundown: Perplexity just launched a new version of its in-house AI model Sonar, built on Llama 3.3 70B and delivering responses at 1,200 tokens per second while still matching or exceeding top models in accuracy at a fraction of the price.

The details:

• Sonar achieves 10x faster responses than competitors like Gemini 2.0 Flash, with Cerebras inference infrastructure enabling near-instant answer generation.

• In tests, Sonar outperformed GPT-4o and Claude 3.5 Sonnet in user satisfaction, factuality, world knowledge, and other benchmarks.

• All Perplexity Pro subscribers now have access to Sonar as their default model, with API access coming soon using the same architecture.

• Perplexity CEO Aravind Srinivas also teased Voice Mode, saying it will be ‘the only product’ that reliably gives real-time voice answers and information for free.

Why it matters: Perplexity continues to pump out major updates — rolling out this speedy new model just 3 weeks after Sonar’s reveal. With ultra-fast response speeds combined with reliable and factual performance topping some of the best models in the industry, Perplexity is making a serious push for a broader chunk of the AI market.

AI TRAINING

📈 Master pre-market analysis with AI insights

The Rundown: ChatGPT's Tasks feature can automate your morning market research by delivering a comprehensive analysis of market sentiment and technical indicators before the market opens.

Step-by-step:

1. Head to ChatGPT and select “GPT-4o with Scheduled Tasks” in the model dropdown menu.

2. Prompt it to set up a core market performance analysis for 6:30 AM EST.

3. Request news and event impact analysis to understand market drivers.

4. Ask for a risk assessment and specific trade opportunities.

Pro tip: Always validate AI-generated trade ideas with your own technical analysis since it might hallucinate. The Rundown University members can access the top prompts here.

PRESENTED BY GALILEO

📖 The essential guide to mastering AI agents

The Rundown: Galileo’s latest eBook is a 100-page deep dive into the world of AI agents — offering practical strategies for selecting, optimizing, and scaling them with confidence.

This comprehensive guide covers:

• How to select the right framework based on your goals

• Best practices for tuning and evaluating agent performance

• How to spot and fix failure points before they become major issues

Click here to get your free copy.

YOUTUBE

🎥 YouTube supercharges creators with AI expansion

Image source: YouTube

The Rundown: YouTube CEO Neal Mohan just unveiled major AI upgrades across the platform, including automated dubbing, expanded creator tools, and a new age detection technology — marking its biggest push yet into AI-powered content creation and moderation.

The details:

• YouTube is expanding its AI detection pilot, giving high-profile creators and artists new tools to ID and control AI content that uses their likeness or voice.

• Auto-dubbing expands to all monetized creators, with YouTube reporting that translated videos generated over 40% of watch time from dubbed versions.

• An AI age estimation system that uses machine learning to detect viewer age ranges and customize content preferences and safety features is rolling out.

• Dream Screen and Dream Track, YouTube's AI generation tools for Shorts, will integrate Google’s Veo 2 for enhanced background and music generation.

Why it matters: YouTube is leveraging AI across all platform areas — which is a win for creators and consumers alike. Plus, with features like auto-dubbing and AI generation tools becoming more widely available, users can streamline the content creation process and get their videos in front of a wider global audience.

QUICK HITS

🛠️ Trending AI Tools

• 🗣️ Zonos-v0.1 - Open-source, real-time TTS models with voice cloning

• 🎥 Luma Ray2 - New image-to-video with enhanced realism and motion

• ⚙️ Cursor - Coding agent with new custom tool integration and upgraded semantic search

• ⚒️ Readdy - Generate professional designs in seconds with ready-to-use code

💼 AI Job Opportunities

• 🎯 The Rundown - Marketing/Media Buyer

• 📦 Coreweave - Warehouse Associate

• 🛠️ Shield AI - Staff Engineer, DevOps

• 💼 Databricks - Startup Account Executive

📰 Everything else in AI today

UC Berkeley researchers unveiled DeepScaleR, a new open-source model that surpasses OpenAI's o1-Preview in mathematical reasoning despite its tiny 1.5B parameter size.

Sam Altman commented on Elon Musk’s offer at the AI Action Summit, calling Musk ‘insecure’ and ‘unhappy’ and saying the antics are to ‘slow us down.’

Apple is reportedly partnering with Alibaba to bring Apple Intelligence to China after previously exploring deals with DeepSeek, Baidu, and ByteDance.

A new study from the Center for AI Safety revealed that LLMs develop internal value systems as they scale, with implications like valuing certain human lives differently and showing resistance to value changes.

Alphabet, OpenAI, Roblox, and Discord launched ROOST, a $27M initiative to develop free, open-source AI tools to combat online child exploitation and promote digital safety.

New BBC research found that major AI chatbots like ChatGPT and Gemini produced significant inaccuracies in over half of the news summaries tested.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Thursday at 4 PM EST to learn how to build your own AI-powered personal assistant to streamline your daily tasks with Dr. Alvaro Cintas, The Rundown’s AI professor.

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. The next chapter of the Elon vs OpenAI feud just arrived, with Musk offering $97B to control his former company.

With billions at stake and egos colliding, Silicon Valley's ultimate power struggle continues to create reality-TV-level drama.

In today’s AI rundown:

• Musk launches $97B bid to retake OpenAI

• AI joins Super Bowl prime time with major commercials

• Transform your ideas into songs with AI

• ByteDance unveils Goku AI image and video creation

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

ELON MUSK & OPENAI

💰 Musk launches $97B bid to retake OpenAI

Image source: Grok / The Rundown

The Rundown: Elon Musk and a group of investors just made an unsolicited $97.4B offer to acquire control of OpenAI's nonprofit arm, escalating tensions once again with his former company and CEO Sam Altman.

The details:

• The bid was submitted by Musk's attorney to OpenAI's board, with backers including xAI, Valor Equity Partners, Baron Capital, and other investment firms.

• The offer comes as OpenAI attempts to transition from nonprofit to for-profit status, with a pending $40B investment from SoftBank at a $260B valuation.

• Musk said he aims to return OpenAI to its open-source roots and promised to match or exceed any competing bids for control of the organization.

• Altman responded dismissively on X, offering to "buy Twitter for $9.74B" instead, leading Musk to call the CEO a ‘swindler.’

Why it matters: The drama never ceases between two of the biggest figures in the tech world, but it's no surprise to see Altman rebuff the offer after Musk’s lawsuits and prodding. With both heavily involved in the U.S. government’s tech push, this likely isn’t the last we’ll see of Musk’s vendetta against the company he helped create.

TOGETHER WITH TURING

📘 The hidden costs of poor LLM strategy

The Rundown: Turing’s white paper breaks down why misaligned training, weak evaluation, and poor optimization create the biggest ROI hurdles for LLMs—not just the model itself.

Inside, you’ll learn how to:

• Avoid hidden costs from ineffective LLM training

• Improve model accuracy and long-term efficiency

• Build a scalable LLM strategy that delivers real business impact

Download the white paper now and learn how to get the most from your AI investments.

THE SUPER BOWL AND AI

🏈 AI joins Super Bowl prime time with major commercials

Image source: OpenAI

The Rundown: AI played a prominent role in Super Bowl LIX’s commercial lineup, with OpenAI, Google, Meta, and others shelling out for advertising slots during the big game with mixed audience results.

The details:

• OpenAI made its SB debut with an artistic black-and-white spot that positioned ChatGPT alongside other historical innovations, such as electricity and space travel.

• Google featured Gemini Live helping a father balance job hunting and parenting, with an earlier spot axed after backlash for incorrect cheese facts.

• Meta showcased its AI-powered Ray-Ban glasses, with Chris Hemsworth and Chris Pratt utilizing features like video recording and its multimodal assistant.

• Other AI products advertised included Salesforce’s Agentforce autonomous agent platform and GoDaddy’s new Airo website creation tool.

Why it matters: AI was everywhere during this year’s Super Bowl, with the usual big players all getting in on the action (with mixed results) on USA Today’s ad rankings. OpenAI’s presence was the biggest wildcard — with the company’s first major spot feeling like a key moment in its transition from a startup into a universally known brand.

AI TRAINING

🎵 Transform your ideas into songs with AI

The Rundown: Suno lets you create original music that matches your creative vision with AI (no musical experience needed).

Step-by-step:

1. Access Suno, select “Create” in the menu options, and enable custom mode.

2. Generate lyrics using either Classic (structured) or ReMi (creative) models.

3. For the style, start with genre tags and add specific instruments, effects, and moods.

4. Once your song is created, expand it using Cover, Persona, or Extend features.

Pro tip: We did an extensive workshop with Sara Adkins, machine learning engineer and artist in residence at Suno, which you can access here (alongside an exclusive Rundown University member free code for a 1-month Premier Plan).

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

🎥 ByteDance unveils Goku AI image and video creation

Image source: ByteDance

The Rundown: ByteDance and the University of Hong Kong just introduced Goku and Goku+, a family of AI models that bridge image and video generation, showing top performance across multiple visual benchmarks and enabling hyperrealistic commercial content for advertising and marketing.

The details:

• Goku achieves top performance on major benchmarks, setting records for both image and video quality with a unified architecture to handle both tasks.

• An advanced "rectified flow" technique enables seamless transitions between images and videos, with the system trained on 160M images and 36M videos.

• An enhanced Goku+ specifically targets advertising and marketing needs, with the ability to create photorealistic human avatars and product demos.

• The + platform includes specialized tools for turning product photos into video clips and creating realistic human-product interactions for commercial content.

Why it matters: Goku’s ability to create extended, high-quality commercial content and product demos shows that AI video production is reaching a new level of sophistication. It is transforming how marketing content is created and consumed, but it is also making it harder than ever to distinguish AI generations from reality.

See Goku’s impressive demo generations here.

QUICK HITS

🛠️ Trending AI Tools

• 💬 Krea Chat - Use natural language commands to create and edit images

• 🤖 Mistral Small 3 - Open-source, 24B model with enhanced speed and accuracy

• 🗣️ Talo - Real-time voice translator for video calls

• 📝 Tana - Put your notes to work with voice and AI

💼 AI Job Opportunities

• 🤖 The Rundown - Writer (Robotics/Tech)

• 📣 Databricks - Field Marketing Manager, Central Europe

• 💻 Coreweave - Data Center Technician

• 👔 Curai - Manager, Provider Operations

📰 Everything else in AI today

OpenAI will reportedly finalize the design for its first generation of in-house AI chips this year and plan to work with TSMC on the initial fabrication.

Zyphra launched Zonos-v0.1 beta, featuring two open-source text-to-speech models with real-time voice cloning capabilities and competitive pricing and quality to rivals.

Anthropic published its Economic Index, a new study tracking AI’s labor market impact — finding that AI usage primarily augments rather than automates work.

Luma AI launched new image-to-video capabilities for its next-gen Ray2 model, showcasing impressive realism and natural motion.

French President Emmanuel Macron unveiled plans for €109B in AI investments ahead of the Paris AI Action summit, including a massive UAE-backed datacenter campus and a €20B commitment from Brookfield to develop infrastructure.

Saudi Arabia pledged a new $1.5B investment into AI inference startup Groq, marking one of the largest single-country commitments to specialized AI chip development.

Sam Altman posted a blog detailing exponential cost reductions in AI computing, predicting widespread AI agent deployment that will reshape economic productivity over the next decade.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Thursday at 4 PM EST to learn how to build your own AI-powered personal assistant to streamline your daily tasks with Dr. Alvaro Cintas, The Rundown’s AI professor.

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. While chatbots previously stumbled over basic arithmetic, DeepMind just announced that its AlphaGeometry2 outperforms math olympiads.

With AI about to reach superhuman math capabilities, are the next wave of models about to unlock a whole new realm of scientific discovery?

In today’s AI rundown:

• DeepMind AI surpasses math olympiads

• Sam Altman speaks on GPT-5

• Run CLI commands from your prompts

• Apple's lamp robot makes expressiveness key to home AI

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

GOOGLE DEEPMIND

🥇 DeepMind AI surpasses math olympiads

Image source: Ideogram / The Rundown

The Rundown: Google DeepMind just introduced AlphaGeometry2, a new version of its math-focused AI model that solved 84% of International Mathematical Olympiad geometry problems from the past 25 years — surpassing the average gold medalist performance.

The details:

• The system combines a Gemini model with a symbolic engine to tackle complex geometry problems requiring rigorous proofs and deductive reasoning.

• AlphaGeometry2 solved 42 out of 50 problems to surpass the average gold medalist score of 40.9, a massive improvement from its predecessor's 54% solve rate.

• The model generated over 300M synthetic theorems and proofs of increasing difficulty for training, featuring a larger and more diverse set than AG1.

Why it matters: Math has typically been one of the areas that language models seem to struggle with (sometimes in simple and comical fashions). Still, DeepMind is quickly cracking the code to unlock systems tackling super-complex problems. This can also play a key role in accelerating other math-heavy scientific areas like physics.

TOGETHER WITH VANTA

🛡️ Live event: AI & security maturity

The Rundown: Join John Hammond — cybersecurity researcher, practitioner, and content creator — and Matt Cooper, Vanta’s Director of GRC, for a fireside chat on AI, security maturity, and the top security risks in 2025.

What you’ll gain:

• A deep dive into 2025’s top cyber risks, including the impact of AI

• Actionable insights to refine your security priorities

• Strategies tailored to your organization’s security maturity level

• A live Q&A at the end

Save your spot and get ready to future-proof your approach to cybersecurity.

OPENAI

🎤 Sam Altman speaks on GPT-5

Image source: UTokyo Center for Global Education

The Rundown: OpenAI CEO Sam Altman and CPO Kevin Weil spoke at the University of Tokyo Center for Global Education, handling questions on the future of AI, upcoming model advances like GPT-5, the company’s new Stargate Project, and more.

The details:

• The CEO said the $500B Stargate Project will allow the next class of models to start developing the first ‘new bits of scientific knowledge.’

• He also revealed that OpenAI’s internal model ranks as the 50th-best programmer in the world and could reach the No. 1 spot by the end of the year.

• Altman expects OpenAI ‘to move in the direction of open source,’ saying that ‘society is willing to take the trade-offs’ that open models bring.

• Altman also compared AI’s capability increases to ‘trying to outrun the calculator,’ saying the tech will be better in every general domain than humans.

Why it matters: Nobody in the business is better than Altman at making mind-bending statements about the future of AI — and this talk had no shortage of them. But if his confidence regarding the coming AI acceleration is any indication, we could see some wild paradigm shifts, like new scientific discoveries, coming sooner rather than later.

AI TRAINING PRESENTED BY PROJECT IDX

✨ Run CLI commands from your prompts

The Rundown: Project IDX is your AI-enabled development environment in the cloud with code assistance from Gemini.

Step-by-step:

1. Log into Project IDX with your Google account.

2. Create a project from the templates dashboard.

3. Click the Gemini icon at the bottom of the workspace or press Cmd+Shift+Space (Ctrl+Shift+Space on ChromeOS, Windows, or Linux).

4. Select “Interactive Chat with Gemini” and pass it a prompt to update your project configuration files and code.

Pro tip: Project IDX offers a streamlined development experience by allowing you to execute terminal commands directly from the user interface.

APPLE

🤖 Apple's lamp robot makes expressiveness key to home AI

Image source: Apple

The Rundown: Apple researchers introduced ELEGNT, a framework aiming to showcase how natural movements can make home robots more engaging and relatable — using an expressive robotic lamp inspired by Pixar's Luxo Jr for demonstration.

The details:

• Apple's prototype combines basic functionality with movements that convey emotions and intentions, like "looking" out a window when discussing weather.

• The robot integrates Siri's voice capabilities while using its movable head and arm to create more natural interactions through gestures and positioning.

• Testing revealed that expressive movements, like nodding or showing curiosity, significantly improve comfort and engagement compared to static responses.

Why it matters: Tech companies are racing to bring robots into our homes, and while many have featured the typical humanoid builds, Apple's research suggests that success may depend on both advanced capabilities and creating devices that can interact in ways that feel more natural and emotionally resonant to users.

QUICK HITS

🛠️ Trending AI Tools

• 🇫🇷 Le Chat - Mistral’s revamped AI assistant platform with 10x response speed and new iOS and Android apps

• 🤖 GitHub Copilot - Microsoft’s coding assistant with new agentic features

• 🎥 Pikadditions - New video-to-video feature that enables users to integrate any subject or object into existing footage

• 👅 Tough Tongue - Multimodal AI agent for navigating difficult conversations

💼 AI Job Opportunities

• 🎯 The Rundown - Marketing/Media Buyer

• 💬 Hippocratic AI - Conversation Designer

• 🚀 Palantir Technologies - Forward Deployed Software Engineer

• 🤝 Databricks - Recruiter

📰 Everything else in AI today

Ilya Sutskever’s SSI is reportedly in talks for new fundraising at a $20B valuation, a 4x increase from September’s round despite no public product or revenue.

OpenAI is establishing a new office in Munich, citing the country’s leading position in European AI adoption with the highest amount of ChatGPT users, paying subscribers, and API developers outside of North America.

OpenAI co-founder John Schulmann is reportedly joining former OpenAI CTO Mira Murati’s new startup after leaving Anthropic after just five months.

Perplexity announced ‘The Million Dollar Question,’ incentivizing users to use the platform and ask questions during the Super Bowl for a chance at a $1M prize.

Over 2,000 artists signed an open letter calling for the cancellation of ‘Augmented Intelligence,’ an upcoming AI art auction at Christie’s — arguing the models use copyrighted work in training.

Krea officially launched its previously teased Chat tool in open beta, allowing users to generate and edit images via a natural language chat interface.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Thursday at 4 PM EST to learn how to build your own AI-powered personal assistant to streamline your daily tasks with Dr. Alvaro Cintas, The Rundown’s AI professor.

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. DeepSeek had the AI world in a tailspin weeks ago, but another challenger from across the Atlantic sped into the AI assistant race in a big way.

This time, it’s not from China but from France, by the EU’s largest AI startup: Mistral. With blazing-fast responses, new features, and flexible deployment options, is the revamped ‘le Chat’ platform the next rival to steal the spotlight from Silicon Valley?

In today’s AI rundown:

• Mistral revamps le Chat assistant

• GitHub powers up Copilot with agentic features

• Outreach automation with ChatGPT

• OpenAI co-founder exits Anthropic

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

MISTRAL

🇫🇷 Mistral revamps le Chat assistant

Image source: Mistral

The Rundown: French AI lab Mistral just unveiled new iOS and Android mobile apps for its ‘le Chat’ assistant, alongside a major platform update featuring enhanced response speeds and enterprise deployment options.

The details:

• The app features core capabilities like web search, document processing, code interpreter, and image generation powered by BFL’s Flux Ultra model.

• Mistral also introduced a new ‘Flash Answers’ feature that processes responses at over 10x the speed of competitors like ChatGPT and Claude.

• New pricing tiers include a free plan, a Pro tier at $14.99/month, a Team tier at $24.99/user/month, and an Enterprise option with custom deployment.

• Enterprise customers gain unique deployment flexibility with options for on-premise installation and custom model implementation.

Why it matters: While Mistral’s models aren’t the industry leaders, its focus on speed and flexible deployment options shows how the AI assistant wars are shifting from raw capabilities to practical advantages like performance and enterprise features – and could now have more of the features needed to entice a broader user base.

TOGETHER WITH CONVEYOR

🛡️ Meet your AI security review sidekick

The Rundown: Conveyor just launched Sue, the first AI Agent for Customer Trust — capable of taking on all the tasks that infosec teams dread like sharing NDA-gated documents, answering security questionnaires and everything in between. Perfect for B2B infosec teams sick of manual work.

Sue can do all of this for you:

• Manage and complete customer security requests while communicating across systems and teams

• Complete, reject, or escalate questionnaires based on parameters you set

• Handle any questionnaire format with no manual tagging — even portals

• Personalize your Trust Center using insights from conversation intelligence tools

Learn more and get started with Sue today.

MICROSOFT

🤖 GitHub powers up Copilot with agentic features

Image source: GitHub

The Rundown: GitHub just announced several major updates to its Copilot assistant, including a new agent mode that can self-correct code, ‘Vision’ for image-to-code generations, an upcoming autonomous coding agent, and more.

The details:

• Agent mode enables Copilot to independently iterate on code, fix runtime errors, and suggest terminal commands while inferring additional tasks.

• A new Vision feature lets users generate code directly from inputted images and mockups, with automatic interface and alt text creation.

• Copilot Edits is now generally available in VS Code, supporting multi-file changes with natural text commands and automatic suggested changes.

• GitHub also teased Project Padawan, an autonomous coding agent coming this year that handles GitHub issues and generates fully tested pull requests.

Why it matters: When the AI coding boom first began, many saw it as a simple complement to human developers. But with constantly accelerating capabilities and emerging agentic features, we’re quickly moving from a completion tool to a true coding partner that can handle complex development tasks independently.

AI TRAINING

📧 Outreach automation with ChatGPT

The Rundown: In this tutorial, you’ll learn how to use ChatGPT o3-mini to automatically analyze company data and generate personalized outreach emails at scale using its reasoning and web search capabilities.

Step-by-step:

1. Structure your company data with key details like name, contact info, and recent updates in a clear format for ChatGPT.

2. Define your outreach goals and configure your AI prompt with instructions for tone and structure to generate personalized emails.

3. Use o3-mini's web search to find new prospects and generate detailed emails based on their updates.

Pro tip: Create an automated workflow with Operator that searches for new companies, generates emails, and drafts them directly in your email client. The Rundown University members can access the top prompts here.

PRESENTED BY VAPI

🗣️ Experience enterprise-ready voice AI

The Rundown: Vapi offers an enterprise-grade platform to deploy voice AI agents in minutes, helping handle calls at scale while staying secure — no matter the use case or industry.

YC startups and Fortune 500 companies rely on Vapi for:

• Flexible integration with your own data, APIs, or models

• Scalability to millions of calls at <500ms latency

• LLM guardrails and SOC-2 compliance

Call 1-844-440-8274 or click here to talk to Vapi.

ANTHROPIC

🚪 OpenAI co-founder exits Anthropic

Image source: UC Berkeley

The Rundown: OpenAI co-founder and key ChatGPT architect John Schulman has reportedly departed AI startup Anthropic after just five months, marking yet another significant shift from a top mind in the field.

The details:

• Schulman originally joined Anthropic in August, citing a desire to focus more deeply on AI alignment research and hands-on technical work.

• Schulman previously spent 9 years at OpenAI as part of the founding team and is credited as a key component of creating ChatGPT.

• Neither Schulman nor Anthropic have detailed the reasons behind the unexpected departure.

• Anthropic's chief science officer, Jared Kaplan, expressed support for Schulman's decision to pursue new opportunities in a statement to Bloomberg.

Why it matters: The sands at the top of the AI leaders are constantly shifting. Schulman’s initial move was viewed as a major shift to the safety-focused Anthropic — and while rumored to have an o3-level model behind closed doors, the startup’s pace of shipping (at least consumer-facing) has ground to a halt since 3.5 Sonnet.

QUICK HITS

🛠️ Trending AI Tools

• ⚡ Concierge - AI Assistant that talks to your SaaS apps in real-time to update your CRM, create Jira tickets, summarize Slack, and more*

• 🚀 Gemini 2.0 Pro - Google’s new advanced AI model with 2M token context

• 🧠 Gemini 2.0 Flash Thinking - Google’s latest AI reasoning model

• 🥒 Pickle - Your AI body double for zoom calls

_{*Sponsored Listing}

💼 AI Job Opportunities

• 🤖 The Rundown - Writer (Robotics/Tech)

• 🎨 Luma AI - Brand Designer

• 💻 Shield AI - Software Simulation Engineer

• 🤝 DeepL - Customer Success Manager

📰 Everything else in AI today

OpenAI is initiating a nationwide search for data center locations across 16 U.S. states to expand its $500B Stargate project beyond Texas.

U.S. bipartisan House lawmakers introduced legislation prohibiting Chinese AI app DeepSeek from being allowed on federal devices, citing national security concerns.

Rideshare giant Lyft is partnering with Anthropic to deploy Claude-powered AI tools across its platform for customer service, product testing, and more.

Google announced that AI-edited images created in Magic Editor's Reimagine feature on Pixel devices will now be tagged with DeepMind's SynthID watermarking tech.

Pika Labs launched Pikadditions, a new video-to-video feature that enables users to integrate any subject or object into existing footage.

TWO AI introduced SUTRA-R0, a multilingual reasoning model that surpasses DeepSeek-R1 and OpenAI-o1-mini in Indian language benchmarks.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today at 4 PM EST to learn how to leverage OpenAI’s Deep Research tool, access open source alternatives, and create your own for free 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. Google’s latest answer in the AI model race just landed, and it’s bringing a whole family of new models to the scene.

But with OpenAI’s buzzy o3 grabbing headlines and 2.0 Pro’s modest benchmark gains, can these new releases keep Google’s December hype train rolling?

Speaking of o3… OpenAI recently launched ‘Deep Research’ — and our next workshop on Friday will teach you how to leverage this powerful new tool (and take advantage of free alternatives). Learn more and register here.

In today’s AI rundown:

• Google rolls out Gemini 2.0 lineup with Pro

• Nvidia's AI teaches robots to move like athletes

• How to research and validate business ideas with AI

• OpenAI signals hardware push with trademark filing

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

GOOGLE

✨ Google rolls out Gemini 2.0 lineup with Pro

Image source: Google

The Rundown: Google just unveiled several new AI models in its Gemini 2.0 lineup, including the highly anticipated Pro Experimental and the cost-efficient Flash and Flash Lite, which also makes its Flash Thinking reasoning model available to all app users.

The details:

• 2.0 Pro Exp. features a massive 2M token context window and excels at coding tasks, with enhanced capabilities for complex prompts and world knowledge.

• A new budget-friendly 2.0 Flash-Lite model delivers better performance than 1.5 Flash while maintaining the same speed and pricing.

• The 2.0 Flash Thinking Experimental reasoning model is now freely available in the Gemini app, showing users step-by-step thought processes in real time.

• All new models feature multimodal input capabilities, with outputs like image generation and text-to-speech planned for release in the coming months.

Why it matters: Google has officially made the leap many were waiting for with its flagship 2.0 Pro model — but unlike the high-powered December releases that were major steps up on the competition, 2.0 Pros benchmarks look a bit underwhelming compared to both 1.5 Pro and the current hype surrounding OpenAI’s latest releases.

TOGETHER WITH SANA

🤝 Bring AI to every team with Sana

The Rundown: Sana AI turns your organization’s knowledge into powerful, secure AI agents — enabling you to launch game-changing AI across every team in minutes, not months.

With Sana, you can:

• Deploy custom AI agents for any role without writing a single line of code

• Unify AI across your organization and create a consistent user experience

• Seamlessly Integrate into existing tools and data sources with enterprise-ready security

Learn more about Sana AI and drive enterprise-wide innovation today.

NVIDIA

🏀 Nvidia's AI teaches robots to move like athletes

Image source: Nvidia

The Rundown: Nvidia and Carnegie Mellon researchers just introduced ASAP, an AI framework that lets humanoid robots learn complex movements from simulations — enabling the replication of iconic celebrations and moves from professional athletes.

The details:

• The system works in two stages: initial training in simulation followed by a specialized neural network that adapts movements for real-world physics.

• Unitree G1 robots demonstrated complex motions in testing, including recreating moves from athletes like LeBron James and Cristiano Ronaldo.

• The framework reduced motion errors by 53% compared to existing methods — a major advance in bridging the gap between virtual and physical training.

• Hardware limitations remain challenging, with two test robots suffering damage from overheated motors during high-intensity movements.

Why it matters: The acceleration in robotic movement capabilities over the last year has been mind-blowing—and what better way to show it than through some of the world's most iconic celebrations? As training gets faster and more efficient, the next step may be for robots to take the actual field in leagues of their own.

Check out video footage of the movements here.

AI TRAINING

🔍 Research and validate business ideas with AI

The Rundown: In this tutorial, you’ll learn how to use Perplexity with the new DeepSeek model to rapidly validate business ideas by analyzing market opportunities, competition, and viability.

Step-by-step:

1. Enable Perplexity Pro and DeepSeek reasoning (5 free daily searches for new users).

2. Provide your business idea and ask to analyze market fundamentals with targeted prompts.

3. Request to research competition and revenue models and create your validation roadmap using its insights.

Pro tip: Use your free DeepSeek searches strategically - save them for complex queries that require advanced reasoning rather than basic market data. The Rundown University members can also access our pre-written prompts 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.

OPENAI

📱 OpenAI signals hardware push with trademark filing

Image source: Grok / The Rundown

The Rundown: OpenAI just filed a trademark application covering everything from humanoid robots to AI wearables, hinting at ambitions to move into physical products through partnerships with hardware veterans like former Apple designer Jony Ive.

The details:

• The application includes smart jewelry, VR/AR headsets, wearables for ‘AI-assisted interaction,’ smartwatches, and more.

• Also listed are ‘user-programmable humanoid robots’ and robots with ‘communication and learning functions for assisting and entertaining people.’

• OpenAI has frequently been linked to Jony Ive, with Sam Altman reiterating that he hopes to create an AI-first phone ‘in partnership’ with him last week.

• The company recently began rebuilding its robotics team, with Figure AI also abruptly ending its collaboration agreement with OpenAI this week.

Why it matters: While competing in consumer hardware would’ve felt insane a few years ago, Apple’s turbulent rollout of AI and unprecedented growth of OpenAI could offer a new window of opportunity. Altman has also said that AI advances warrant a “new kind of hardware” — and maybe humanoids are that final form he envisions.

QUICK HITS

🛠️ Trending AI Tools

• 🎉 Invites - Apple’s AI-powered party planning tool

• ⚡️ Livekit - New open-source turn detection model for natural voice AI

• ⚙️ a0 Dev - Generate full mobile apps with AI in minutes

• 💬 Chatbase - AI agents for magical customer experiences

💼 AI Job Opportunities

• 🎯 The Rundown - Marketing/Media Buyer

• 🧵 Figure AI - Senior Soft Goods Engineer, TeleOp

• 🤝 Databricks - Business Development Representative

• 🚀 Luma AI - Growth Marketer

📰 Everything else in AI today

Free 1-hour AI workshop: Get proficient in prompting, learn solid AI use cases, and get confident using AI in just one hour with Section. Enroll for free.*

Google revised its AI ethics principles to remove restrictions on the use of the technology for weapons and surveillance applications.

OpenAI shared a demo of an automated sales agent system during an event in Tokyo, which has the ability to handle tasks like enterprise lead qualification and meeting scheduling.

Amazon scheduled a hardware event for Feb. 26 in New York, where it is expected to unveil its long-awaited AI-enhanced Alexa overhaul.

Enterprise software giant Workday announced plans to cut 1,750 jobs or 8.5% of its workforce as part of an AI-driven restructuring plan.

MIT researchers unveiled SySTeC, a tool that speeds up AI computations by automatically eliminating redundant calculations, achieving up to 30x speed increases.

^{*Sponsored Listing}

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Friday at 4 PM EST to learn how to leverage OpenAI’s Deep Research tool, access open source alternatives, and create your own for free 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. Researchers from TikTok owner ByteDance just unveiled OmniHuman-1, showing off the most insane AI model yet for generating videos indistinguishable from real footage.

They can clone any human into a video with just a single photo and audio input. But is society ready for what this means for the future of digital content?

In today’s AI rundown:

• ByteDance’s AI-human model breakthrough

• Apple introduces AI-powered party planner

• Build a personal AI automation tool

• AI creates a database to help detect cancer earlier

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

BYTEDANCE

🎭 ByteDance’s AI-human avatar model breakthrough

Image source: ByteDance

The Rundown: Researchers from TikTok owner ByteDance just unveiled OmniHuman-1, an AI system that can generate some of the most realistic deepfake videos we’ve ever seen, all from just a single reference image and audio input.

The details:

• The system can create convincing videos of any length and style, with adjustable body proportions and aspect ratios.

• It handles diverse inputs from cartoons to challenging human poses while maintaining style-specific motion characteristics.

• It’s trained on 19,000 hours of video and can even modify movements in existing footage.

• Despite 10 U.S. states enacting laws against AI impersonation, detection and regulation remain major challenges.

Why it matters: ByteDance hasn't publicly released OmniHuman-1, but the demos have officially erased the line between real and AI-generated. As similar powerful systems inevitably become available, society faces an urgent challenge: verifying what's real in a world where anyone can create perfectly fake videos.

TOGETHER WITH DYNAMIQ

🤖 Build AI agents in just hours

The Rundown: Dynamiq empowers individuals to break free from traditional no-code tool constraints and build multi-agent systems with a fully customizable low-code interface.

With Dynamiq, you can:

• Build agents at zero cost with our free plan (no credit card needed)

• Avoid tedious setups with pre-integrated connections to the top LLMs

• Jumpstart your projects using ready-to-go agent templates, fully customizable using Python code

Start building for free.

APPLE

🍎 Apple introduces AI-powered party planner

Image source: Apple

The Rundown: Apple just released Invites, a new AI-powered event planning app that integrates Apple Intelligence with multiple Apple Services to create custom invitations and manage events.

The details:

• The app uses AI to generate custom images and text for invitations through Image Playground and Apple Intelligence Writing Tools.

• It also integrates multiple Apple services (Photos, Music, Maps, Weather) into a single event portal.

• Unlike most Apple services, it's accessible to non-Apple users for RSVPs and photo sharing.

• While free to download in the app store, this marks Apple's first AI-powered standalone app, suggesting a shift in their AI strategy.

Why it matters: While competitors race to build powerful models, Apple takes a different approach by integrating AI into focused, practical apps. The company is still finding its footing after a rocky start with Apple Intelligence, but its track record of perfecting features through iteration might be exactly what's needed.

AI TRAINING

⚡️Build a personal AI automation tool

The Rundown: In this tutorial, you’ll learn how to build your own AI automation running locally and privately using N8N's automation platform and Ollama's language models, such as DeepSeek.

Step-by-step:

1. Install Docker Desktop and Ollama, then set up N8N using Docker commands found here.

2. Download a language model through Ollama (7B or 14B size recommended).

3. Create a new N8N workflow with a chat trigger and AI Agent node.

4. Select your Ollama model and add Window Buffer Memory for conversation context, then start chatting!

Pro tip: Use "host.docker.internal:11434" as your base URL in N8N to ensure smooth communication between N8N and Ollama.

PRESENTED BY GALILEO

🤖 Your path to powerful AI agents

The Rundown: Galileo’s new eBook brings you a 100-page roadmap to creating reliable AI agents — complete with strategies for navigating issues like framework selection, performance tuning, and avoiding common pitfalls.

With this guide, you’ll learn how to:

• Select the right agentic framework for your needs

• Evaluate and improve agent performance with proven techniques

• Identify and resolve failure points before they impact production

Get your copy and start building stronger AI agents today.

AI RESEARCH

🔬 AI creates a database to help detect cancer earlier

Image source: John Hopkins University

The Rundown: Johns Hopkins researchers just created AbdomenAtlas, an AI-powered dataset of 45,000 3D CT scans with 142 annotated anatomical structures, dramatically accelerating medical image analysis for early cancer detection.

The details:

• The dataset is 36 times (!) larger than its closest competitor, combining scans from 145 hospitals worldwide.

• Using AI and 12 expert radiologists, the team completed in two years what would have taken humans 2,500 years.

• The system achieved a 500-fold speedup for organ annotation and 10-fold for tumor identification.

• The team plans to release AbdomenAtlas publicly and continues adding more scans, organs, and tumor data.

Why it matters: AbdomenAtlas could transform early cancer detection by giving AI models much more comprehensive training data. However, even at 45,000 scans, it represents just 0.05% of annual US CT scans — highlighting how early we are in building truly comprehensive medical AI systems.

QUICK HITS

🛠️ Trending AI Tools

• 🎙️ PlayAI Dialog 1.0 - AI text-to-speech model with a focus on ‘ultra-emotional voices’

• 🧑‍💻 Replit iOS App - Turn your ideas into apps with Replit’s new iPhone AI app

• 🧠 OpenAI Kit - A free portal to in-depth research responses from OpenAI’s DeepResearch feature

• 🐼 Earkick - An AI mental health iOS app that’s fully anonymous with no registration or data collection required

💼 AI Job Opportunities

• 🤝 The Rundown - Partnerships Manager

• 💼 UiPath - Enterprise Sales Executive

• 📈 Waymo - Product Data Science Lead, Supply Operations

• 🔎 Databricks - Senior Software Engineer, Search

📰 Everything else in AI today

Free event - 2/18: The AI Valuation Playbook with EY’s global AI leader. Get practical tips for using AI to reach company objectives. RSVP now*

Figure ended its collaboration agreement with OpenAI, hinting at a major breakthrough in end-to-end robot AI to be revealed within 30 days.

Kanye West confirmed he's using AI on his upcoming album 'BULLY,’ comparing the role of technology in music to that of autotune.

LiveKit introduced a new transformer model for more natural AI voice conversations, reducing unintentional interruptions by 85% through improved end-of-turn detection.

Google published its 2024 Responsible AI Progress Report and updated its Frontier Safety Framework, introducing new protocols for managing AI risks and security.

Hugging Face released open-Deep-Research, an open-source alternative to OpenAI's Deep Research, achieving 55% accuracy on the GAIA benchmark with autonomous web navigation capabilities.

Adobe enhanced Acrobat's AI Assistant with contract intelligence features to help users understand complex legal documents and identify key terms.

Snap unveiled a mobile-first AI text-to-image model that can generate high-resolution images in 1.4 seconds on iPhone 16 Pro Max, and it plans to integrate it into Snapchat features.

_{* Sponsored listing}

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Friday at 4 PM EST to learn how to leverage OpenAI’s Deep Research tool, access open source alternatives, and create your own for free 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. SoftBank just announced a massive $3B annual commitment to OpenAI alongside 'Cristal Intelligence'’ a joint venture to bring customized AI tools exclusively to Japanese businesses.

With Stargate's $500B data centers and now Cristal Intelligence, are we witnessing Japan's big move in the global AI race?

Community note: We’re hosting a live workshop today at 3 PM EST going in-depth how to build and deploy complete mobile apps with AI. RSVP here.

In today’s AI rundown:

• SoftBank and OpenAI announce 'Cristal Intelligence'

• Anthropic challenges hackers to break its AI

• Automate meeting prep with AI Agents

• EU invests $56M to build open-source AI rival

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

SOFTBANK

🇯🇵 SoftBank and OpenAI announce 'Cristal Intelligence'

Image source: SoftBank

The Rundown: SoftBank just announced a $3 billion annual commitment to OpenAI's technology while launching ‘Cristal Intelligence,’ a joint venture to bring customized OpenAI tools exclusively to Japanese businesses.

The details:

• SoftBank will spend $3B annually on OpenAI technology across its subsidiaries, including Arm and PayPay.

• Cristal Intelligence will offer a specialized business version of ChatGPT and OpenAI's API, which is exclusively available in Japan.

• The joint venture aims to provide secure, customized enterprise integrations, potentially competing with traditional consulting firms.

• This partnership follows SoftBank's involvement in "Stargate," a $500B data center initiative with OpenAI and Oracle.

Why it matters: This massive investment ($3B per YEAR) signals Japan's serious push into the global AI race through SoftBank's growing alliance with OpenAI. With their recent Stargate partnership and now Cristal Intelligence, SoftBank is going all in on building the infrastructure to dominate AI deployment across Asia.

TOGETHER WITH HUBSPOT

💡 Save 10+ hours per week with an AI personal assistant

The Rundown: HubSpot’s free ‘AI Task Delegation’ playbook helps turn artificial intelligence into your own personal assistant, enabling you to delegate routine work effectively, boost productivity, and reclaim over 10 hours each week.

Inside, you’ll find:

• The exact templates used by top performers to delegate 80% of routine work to AI

• An all-in-one Command Center to integrate and manage your AI tools

• A step-by-step implementation guide to mastering AI delegation in under 60 minutes

• A built-in ROI calculator for time savings, advanced ChatGPT prompts, and more

Download your Free AI Assistant Kit now.

ANTHROPIC

🔒 Anthropic challenges hackers to break its AI

Image source: The Rundown / Ideogram

The Rundown: Anthropic just unveiled Constitutional Classifiers, a new AI safety system with promising results, and is inviting the public to help stress-test it after surviving over 3,000 hours of unsuccessful bug bounty attempts.

The details:

• The system uses AI to generate training data in multiple languages and writing styles, helping it catch diverse jailbreak attempts.

• In testing against 10,000 advanced jailbreak attempts, it blocked 95.6% of attacks, compared to just 14% for unprotected Claude.

• 183 bug bounty hunters spent over 3,000 hours trying to break the system for a $15,000 reward, but none succeeded in fully jailbreaking it.

• Anthropic is inviting the public to test the system until February 10.

Why it matters: As AI models become increasingly powerful, it is crucial to prevent them from being manipulated to cause harm. While most companies rely mainly on training for AI safety, Anthropic's new approach of using AI to catch jailbreaks shows promise, and opening it to public testing flaunts their commitment to making AI safer.

AI TRAINING

🤖 Automate meeting prep with AI Agents

The Rundown: Lindy AI allows anyone to create AI agents with no code, such as a ‘Pre-meeting Prep Agent‘ that automatically researches attendees, gathers email history, and sends you a comprehensive pre-meeting brief before every meeting.

Step-by-step:

1. Visit Lindy AI and create a new automation from scratch, or select a starting template.

2. Set up ‘Calendar Event Started trigger’ (customize timing for pre-meeting briefs).

3. Add conditions to check for external attendees and configure research actions, e.g., internet search or email history.

4. Set up email brief delivery with all the gathered information.

5. Your AI agent can send briefs 15 minutes pre-meeting, saving you time to prep in between tight windows.

Pro tip: We did an extensive workshop with Flo Crivello, CEO of Lindy AI, which you can access here (alongside an exclusive Rundown University member discount rate).

PRESENTED BY SYNCLY

🧠 Your feedback-to-insights engine

The Rundown: Syncly (YC W23) is a customer feedback intelligence platform that processes customer feedback—giving you actionable insights to reduce churn, refine products, and improve customer experiences at scale.

Syncly helps you:

• Quickly identify and address issues impacting customer experience

• Stay ahead of churn by spotting and engaging at-risk customers before they leave

• Employ self-training AI for deeper insights with less hands-on effort

• Leverage analytics to spot new opportunities and operational improvements

Book a free trial now and uncover your customer intelligence today.

EUROPEAN UNION

🇪🇺 EU invests $56M to build open-source AI rival

Image source: The Rundown / Ideogram

The Rundown: The European Union has announced a $56M investment to develop OpenEuroLLM, a new open-source large language model that works with all 30 European languages.

The details:

• The project will leverage EU supercomputers like Spain's Mare Nostrum and Italy's Leonardo.

• While $56M is tiny compared to OpenAI's reported $40B raise, it's 10x what DeepSeek claimed to have spent on their breakthrough model.

• The initiative promises fully open models, software, and data that can be fine-tuned for specific sectors like healthcare and banking.

• The goal is to create an open-source LLM that European companies and governments can build upon, with EU values "baked in."

Why it matters: After years of scrutiny focusing on AI regulation, the EU is finally stepping into the AI race (…sort of). While $56M is puny compared to major AI giants investing billions, the EU may succeed in specialized, industry-specific AI, especially following DeepSeek's recent success with low-budget open-source models.

QUICK HITS

🛠️ Trending AI Tools

• 🧠 OpenAI Deep Research - A new ChatGPT Pro plan feature that does extensive web research and delivers detailed reports with citations

• 🕵️ Geospy AI - Finds the location from any photo by analyzing its pixels

• 📝 Tana - An AI-native workspace that turns notes into work tasks

• 🔥 Open Deep Research - An open-source alternative to Gemini and OpenAI Deep Research

💼 AI Job Opportunities

• 🤖 The Rundown - Writer (Robotics/Tech)

• 🌏 Palantir Technologies - Global Mobility Specialist

• 👥 OpenAI - Recruiting Manager, Go To Market & Sales

• 🛠️ Together AI - Summer 2025 Engineering Intern

📰 Everything else in AI today

ARC Prize found OpenAI’s new o3-mini AI model to patch o1 on its ARC-AGI-1 Semi-Private Test Set while being 100x less the cost.

Meta published its Frontier AI Framework, emphasizing its commitment to open-source development while focusing on mitigating cybersecurity and weapon risks.

The Beatles won a Grammy for Best Rock Performance with "Now and Then," an AI-enhanced song that used noise reduction to clean up an old Lennon demo.

OpenAI expanded ChatGPT's WhatsApp integration globally, adding image uploads and voice message support simply by texting 1-800-CHATGPT (1-800-242-8478).

UK researchers developed self-healing asphalt using biomass waste and Google Cloud's AI, potentially solving the country's ‘pothole crisis’.

Microsoft is forming a new unit called the Advanced Planning Unit (APU) within its AI division to study AI's implications for society, health, and work.

COMMUNITY

🎥 Join our next live workshop

Join our next live workshop today at 3 PM EST for a special 90-minute deep dive into building revenue-generating mobile apps using AI with Kevin Hou, Head of Product Engineering at Codeium.

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