Alibaba’s AI video powerhouse

Good morning, AI enthusiasts. China’s Alibaba just released a robust AI video generation suite that ANYONE can download, run, and modify — and early tests suggest it outperforms even OpenAI’s Sora.

While most AI labs keep its best AI locked behind waitlists and paywalls, Chinese labs are beginning to push ahead with open-source alternatives that anyone can use today.

Speaking of AI video, our next workshop is today at 3 PM EST — where you’ll learn create personalized AI video messages at scale with Synthesia’s Kevin Alster. RSVP.

In today’s AI rundown:

• Alibaba’s advanced AI video suite

• Google’s free AI coding assistant

• Find your next SaaS idea using AI research

• Claude plays Pokemon Red live to tell the tale Twitch

• 4 recent AI tools & 4 job opportunities

LATEST DEVELOPMENTS

ALIBABA

🎬 Alibaba’s advanced AI video suite

The Rundown: Alibaba’s Tongyi Lab just released Wan2.1, an open-source suite of powerful video generation models that outperform SOTA open-source and closed models equivalent to Sora on key benchmarks — while generating videos at 2.5x the speed.

The main points:

• Wan2.1-T2V-14B tops the VBench leaderboard, excelling in areas like complex motion dynamics, real-world physics simulation, and text generation.

• All models support text-to-video, image-to-video, and video-to-audio, and are the primary with the power to render text in each English and Chinese.

• Wan’s editing tools include video inpainting and outpainting, multi-image referencing, and the power to take care of existing structures and characters.

• The discharge also includes a light-weight 1.3B version able to running on consumer hardware—it will probably generate a 5-sec 480P clip on RTX 4090 in 4 minutes.

Why it matters: One other day, one other wild open-source release out of China. Wan is a continuation of the accelerating quality we’ve seen from recent launches like Google’s Veo 2 — with telltale AI signs (choppy motion, artifacts, etc.) all but completely eliminated. Between Qwen and Wan, Alibaba is bringing the open-source heat in 2025.

TOGETHER WITH GOOGLE CLOUD

🔮 AI predictions every startup must know

The Rundown: Google Cloud’s Way forward for AI: ‘Perspectives for Startups’ report gathers predictions, real-world applications, and expert strategies from 23 AI leaders like Amin Vahdat (Google Cloud) and Chamath Palihapitiya (Social Capital) to provide help to position your enterprise for achievement in 2025 and beyond.

What’s inside:

• Expert-backed predictions on AI’s role in business and tech

• Lessons from real-world AI adoption at startups

• Practical strategies to leverage AI for a competitive advantage

• Areas investors prioritize for AI startups

Click here to download the report.

GOOGLE

🔧 Google’s free AI coding assistant

The Rundown: Google just launched a free version of Gemini Code Assist for individual developers, offering access to advanced AI-powered coding help with usage limits that dwarf competitors like GitHub Copilot.

The main points:

• Gemini Code Assist is powered by a fine-tuned version of Google’s Gemini 2.0 model optimized specifically for programming tasks.

• The brand new tool provides as much as 180,000 monthly code completions — 90 times greater than GitHub Copilot’s free tier limit of two,000.

• The assistant encompasses a 128,000 token context window, allowing it to process and understand much larger codebases than competitors.

• The free version also integrates with dev environments like Visual Studio Code, GitHub, and JetBrains, with just a private Google account needed.

Why it matters: AI has modified programming perpetually, with powerful free tools driving the most important shift. Google’s latest push with Gemini Code Assist could further disrupt this market dominated by GitHub Copilot—unlocking recent possibilities for developers worldwide.

AI TRAINING

🧠 Find your next SaaS idea using AI research

The Rundown: On this tutorial, you’ll learn use Perplexity’s Deep Research to construct a validated software product, complete with development specifications and launch strategy.

Step-by-step:

1. Visit Perplexity and choose “Deep Research”.

2. Use the prompt: “You’re a Market Research Specialist. Analyze [INDUSTRY] by identifying top 10 problems, severity levels, current solutions, and market size.”

3. Transform your chosen problem right into a product by prompting: “Design a SaaS solution for [problem], including essential features, tech stack, timeline, and revenue streams.”

4. Get implementation details with: “Provide step-by-step coding instructions, including API specifications and deployment configurations.”

Pro tip: In the event you don’t know code, AI coding assistants like Cursor, Windsurf, or Replit Agent can massively provide help to construct your recent idea.

PRESENTED BY UNSTRUCTURED

🙅 Ditch your old data connectors

The Rundown: Join Unstructured for a deep dive into best practices for data connectors for generative AI applications, and see where solutions like Fivetran, Boomi, and AirByte fall short with maintaining wealthy context or keeping your team out of the weeds updating them.

On this live session, you’ll discover:

• Frameworks for evaluating ETL tools across critical features

• Why traditional solutions struggle with messy data sources

• Best practices to maintain engineers focused on innovation — not maintenance

Register to affix or get a recording.

AI RESEARCH

🕹️ Claude plays Pokemon Red live to tell the tale Twitch

The Rundown: Anthropic just debuted “Claude Plays Pokémon” on Twitch, a continuation of the corporate’s research showcasing their recent AI model Claude 3.7 Sonnet attempting to navigate the classic Game Boy game Pokémon Red in real-time.

The main points:

• 3.7 Sonnet made major progress in comparison with its predecessors, defeating three gym leaders — with the unique Sonnet struggling to go away the starting location.

• The livestream shows Claude’s “thought process” on the left while real-time gameplay appears on the fitting, giving viewers insight into the AI’s reasoning.

• Claude has access to a knowledge base to store info, function calling to take actions, and vision capabilities to look at the sport.

• Unlike previous versions, 3.7 Sonnet’s reasoning capabilities help navigate the sport more effectively — planning, adapting, and remembering objectives.

Why it matters: Watching Claude tackle Pokemon is each fascinating and endearing — especially when seeing the thought process behind every move. While we’ll be playing games alongside AI agents in the longer term, Claude’s experiment can be an interesting window right into a future where AIs are observed passively as popular gaming streamers.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 Claude 3.7 Sonnet and Claude Code – Latest reasoning AI and coding agent

• 🧠 QwQ-Max – Alibaba’s open-source reasoning LLM

• 🔊 Luma Labs Video to Audio – Generate synced audio for video generations

• 🔥 EasyGen – A LinkedIn post generator that creates algorithm-optimized posts

💼 AI Job Opportunities

• 🤖 Figure AI – FPGA Design Intern

• 🖥️ Anyscale – Software Engineer, ML Developer Platform

• 💰 Coreweave – Credit and Collections Manager

• 📊 UiPath – GTM Operations Planning Program Manager

📰 Every thing else in AI today

OpenAI rolled out its Deep Research feature to ChatGPT Plus, Team, Edu, and Enterprise tiers, offering 10 queries monthly in comparison with the $200 Pro tier’s 120.

Anthropic is reportedly set to boost a larger-than-planned $3.5B funding round at a $61.5B valuation, with the news coming just days after the discharge of Sonnet 3.7.

DeepSeek is reportedly aiming to push the discharge of its next AI model up from its initial May timeline, looking for to capitalize on its success within the wake of R1.

OpenAI also launched a GPT-4o mini-powered Advanced Voice to free ChatGPT users—promising the identical conversation style because the GPT-4o version for Plus and Pro users.

Microsoft removed usage limits on Copilot’s Voice and Think Deeper features, giving all free users unlimited access. Pro users will retain priority access during peak hours.

Over 1,000 musicians released a silent album to protest the UK’s proposed copyright changes that might allow AI corporations to coach on works without explicit permission.

IBM announced plans to accumulate DataStax, aiming to assist unlock enterprise data for AI applications while expanding its NoSQL database offerings.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today, Wednesday, February twenty sixth at 3 PM EST and learn create personalized AI video messages at scale with Synthesia’s Kevin Alster, Strategic Advisor and former Head of Synthesia Academy.

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

See you soon,

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