AI takes center stage at Davos

Good morning, AI enthusiasts. The world’s elite converged on Davos this week for the World Economic Forum, and there was one topic on everyone’s minds — AI.

The takeaway from the industry’s top leaders? Disruption is coming faster than expected, and the window for each individuals and corporations to adapt is closing fast.

P.S.: We’re sitting down for a Q&A with Jeetu Patel, President and Chief Product Officer at Cisco, and need our community to take part in the motion! Submit any AI questions on the industry, security, future predictions, company workflows, etc., here.

In today’s AI rundown:

• AI takes center stage at WEF in Davos

• ‘Human-centric’ AI startup raises $480M

• Recover AI outputs using multiple selection

• LTX launches audio-to-video generation

• 4 latest AI tools, community workflows, and more

LATEST DEVELOPMENTS

AI & THE WORLD ECONOMIC FORUM

🌎 AI takes center stage at WEF in Davos

The Rundown: The World Economic Forum kicked off this week in Davos, featuring an intense give attention to AI that included discussions with the industry’s top minds, including Anthropic’s Dario Amodei, Google’s Demis Hassabis, and Microsoft’s Satya Nadella.

The small print:

• Amodei took a shot at U.S. President Donald Trump’s recent policy allowing AI chip sales to China, likening it to “selling nuclear weapons to North Korea”.

• Hassabis believes this yr could see AI-driven slowdowns in junior hiring, but that AI tools could ultimately enable more skill creation than traditional paths.

• Nadella said no company can “just coast” within the AI era, stating that big firms that don’t sustain will “get schooled by someone small.”

• Amodei also said we could also be just 6-12 months away from models that do “most, perhaps all” of what software engineers do end-to-end.

Why it matters: Davos talks often stay abstract, but Amodei’s timeline for full AI-powered coding puts a good deadline on the disruption we’re already seeing with tools like Claude Code. Between geopolitical warnings, labor shifts, and software upheaval, the message from AI’s leaders is obvious: the window to adapt is shrinking rapidly.

TOGETHER WITH MONGODB

📈 Startup support that delivers results

The Rundown: Let MongoDB for Startups partner with you as you construct the subsequent big thing with the tools, support, and go-to-market opportunities you must speed up from idea to IPO.

MongoDB for Startups includes:

• Built-in vector search using VoyageAI

• Real-time data streaming

• Baseline credits and technical expertise

Start today totally free.

HUMANS&

🤝 ‘Human-centric’ AI startup raises $480M

The Rundown: A brand new AI startup founded by researchers from Anthropic, xAI, and Google, just raised $480M in seed funding at a $4.48B valuation, pushing to construct AI that helps people collaborate higher over automation and full substitute.

The small print:

• Humans& goals to construct AI that functions like an intelligent group chat, requesting context, storing memory, and coordinating teams.

• Co-founder Andi Peng left Anthropic over its autonomy focus, saying they “loved to focus on how its models churned for 8, 24, 50 hours by themselves.”

• The team includes Google’s seventh worker, Georges Harik, two former xAI researchers who worked on Grok, and Stanford professor Noah Goodman.

• The $480M seed round was led by SV Angel and its co-founder Harik, alongside Nvidia, Jeff Bezos, and Google Ventures.

Why it matters: A $4.48B value for a 3-month old startup is wild, but shows the investor appetite for any team with a frontier lab pedigree. The “human-centric” angle is a shot on the autonomy-driven gains most labs are chasing, though we’ve seen similar framing from the likes of Microsoft AI (and even Anthropic in its own ads).

AI TRAINING

📝 Recover AI outputs using multiple selection

The Rundown: On this tutorial, you’ll learn a universal technique to recover outputs from any AI model at a fraction of the associated fee by minimizing the quantity of unhelpful context you add to the AI’s “brain”.

Step-by-step:

1. Open any chat-based AI and break your idea into Goal, Task, and Next Steps to “minimize unhelpful messages”. The duty might be to interview to construct context

2. Prompt using the Goal→Task→Next Steps pattern, e.g.: “Your job is to generate a logo… Interview me with 5-10 MCQs about my brand, vision to construct context”

3. Reply to Qs with easy decisions (e.g., “A, B, A, C…”). This removes miscommunication and ensures that the AI has a greater understanding

4. Once it generates concepts, prompt to “Generate a 4×4 grid with each concept in a separate grid cell” to save lots of on tokens, then ask for standalone images

Pro tip: If you happen to get great results from this method, export and re-use the context by asking the AI to “extract the context as system instructions in markdown.”

PRESENTED BY MAULT

💻 The true-time governance layer for AI-generated code

The Rundown: Mault is a VS Code extension that applies real-time architectural guardrails to stop drift as code is generated.

Unlike linters or CI checks, Mault operates at the purpose of change:

• Detects structural and config mismatches

• Enforces file structure and naming

• Flags deprecated or unsafe dependencies

• Works alongside AI coding tools

Click here to download the extension.

LIGHTRICKS

🎵 LTX launches audio-to-video generation

The Rundown: AI creative platform Lightricks just released a brand new Audio-to-Video feature, allowing users to start out with voice, music, or sound effects after which generate the video output to accompany it.

The small print:

• The tool syncs motion, lip movements, and camera work on to the uploaded audio, keeping consistency across generations.

• Users can upload files, record directly, or use a built-in text-to-speech tool powered by ElevenLabs’ Scribe V2 model.

• LTX said the model can understand rhythm and beat, making it able to music videos and timing/pacing outputs to align with audio.

• LTX called with the ability to start from audio “the third paradigm in AI video generation,” allowing for more consistent and natural outputs.

Why it matters: Starting with sound opens up brand latest creative pathways — with the power to construct around existing audio assets quite than roughly fitting dialogue onto AI-generated clips. While AI video has already hit such a high bar, tools enabling consistency and customization are the subsequent lever to take models to latest heights.

QUICK HITS

🛠️ Trending AI Tools

• 💡 Skills – Vercel’s open ecosystem for locating and sharing agentic skills

• 🎥 LTX Audio to Video – Generate videos directly from audio inputs

• 👁️ Step3-VL-10B – StepFun’s open-source SOTA vision language model

• 💦 LFM2.5-1.2B-Considering – Liquid AI’s latest on-device reasoning model

📰 The whole lot else in AI today

OpenAI’s Chris Lehane revealed during a chat at Davos that the corporate is “on target” to unveil its Jony Ive-led physical AI device within the second half of 2026.

Node.js creator Ryan Dahl declared the era of humans writing code over, arguing that software engineers still have work, but writing syntax directly is not any longer the job.

Liquid AI released LFM2.5-1.2B-Considering, a reasoning variant sufficiently small to run on smartphones while matching larger rivals on math and problem-solving benchmarks.

OpenAI introduced a brand new age prediction feature on ChatGPT to assist mechanically apply underage-specific safeguards to younger users.

AI inference company Baseten is raising $300M in funding, with half coming from an investment by Nvidia, valuing the startup at $5B.

Anthropic announced a brand new partnership with Teach For All to coach greater than 100k educators across 63 countries on AI tools for the classroom.

COMMUNITY

🤝 Community AI workflows

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

Today’s workflow comes from reader Matt R. in London, UK:

“I used to be making my annual family calendar and located that plenty of my wife’s Apple photos were in HEIC format, which weren’t accepted by Snapfish. I searched for a file converter to get them into JPGs, and was quickly browsing the spate of ‘free’ options that do not let you know that to convert multiple file, you might have handy over your bank card.

I asked Mocha to construct me a HEIC to JPG converter, and 5 minutes later, it was ready and operational!”

How do you utilize AI? Tell us here.

🎓 Highlights: News, Guides & Events

• Read our last AI newsletter: Claude Code sparks AI ‘selfware’ era

• Read our last Tech newsletter: Biotech’s big 3 (and so they’re wild)

• Read our last Robotics newsletter: 1X now has a world model

• Today’s AI tool guide: Recover AI outputs using multiple selection

• RSVP to next workshop @ 4PM EST Thursday: AI Foundations Bootcamp pt.3

See you soon,

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