The Future of Shopping? AI + Actual Humans.

AI has changed how consumers shop by speeding up research. But one thing hasn’t changed: shoppers still trust people more than AI.

Levanta’s new Affiliate 3.0 Consumer Report reveals a major shift in how shoppers blend AI tools with human influence. Consumers use AI to explore options, but when it comes time to buy, they still turn to creators, communities, and real experiences to validate their decisions.

The data shows:

• Only 10% of shoppers buy through AI-recommended links

• 87% discover products through creators, blogs, or communities they trust

• Human sources like reviews and creators rank higher in trust than AI recommendations

The most effective brands are combining AI discovery with authentic human influence to drive measurable conversions.

Affiliate marketing isn’t being replaced by AI, it’s being amplified by it.

Download the full report to see what this means for your brand.

Hey everyone,

I hope you’re all having a great week as we wrap up 2025. It’s been a wild year for tech, hasn’t it? Just when we thought things were quieting down for the holidays, the industry decided to drop some absolute bombshells.

Today:

• Meta Buys AI Agent Startup Manus 

• Nvidia Invests $5 Billion in Intel 

• New Chinese GPU Hits the Market 

• Japan Quadruples Budget for AI & Chips

• Data Centers Turning to Jet Engines for Power

Meta just bought Manus — and it’s basically a bet that “AI agents” are ready for the mainstream

Meta says it’s acquiring Manus, a general-purpose AI agent startup, and plans to operate and sell Manus as a service while also integrating it into both consumer and business products (including Meta AI).

What made me sit up: Meta isn’t pitching this like “another chatbot.” Manus is framed as something closer to a digital employee that can run multi-step work (research + automation) with minimal prompting.

If you’ve been watching the “agent hype vs. agent reality” gap, this is Meta basically saying: we think the gap is closing, and we want distribution on our side when it does.

Why it matters:

• If Meta can bake agent workflows into the places people already live (Messenger / Instagram / WhatsApp / Meta AI), it turns “agents” from a niche power-user thing into a default habit.

It also signals a more aggressive “buy + integrate” posture as the AI platform race tightens.

Nvidia finalized a $5B Intel stake — and that’s a very loaded signal

Nvidia completed the purchase of $5 billion in Intel shares, buying over 214.7 million shares at $23.28 each under an agreement announced back in September.

This reads like more than a financial headline. Intel needs capital (and credibility), while Nvidia benefits from having more strategic leverage across the ecosystem—especially as compute demand keeps warping every hardware roadmap.

The way I’d translate it:

• For Intel: this is a high-profile “vote of confidence” at a moment when the company is trying to rebuild momentum.

• For Nvidia: it’s optionality. Even if nothing dramatic changes tomorrow, Nvidia is quietly positioning itself where it can influence future platform decisions (and partnerships).

Also worth noting: U.S. antitrust agencies cleared the investment (FTC notice earlier in December), which removes at least one obvious near-term speed bump. 

China’s Lisuan G100 family is reportedly shipping — the “domestic GPU” story is getting real (slowly)

Lisuan’s G100-series GPUs have reportedly begun initial deliveries, with early shipments aimed more at enterprise / professional workloads first (think “digital twin” customers), while broader retail timing is still unclear. 

Two details stood out:

• The family includes a gaming-oriented 7G106 and an enterprise-focused 7G105.

• Specs being discussed publicly are ambitious for a first serious swing: 7G106 is described with 12GB GDDR6 on a 192-bit bus, PCIe 4.0, and a custom upscaler (NSRR / NRSS mentioned across coverage).

The bigger point isn’t “this beats Nvidia tomorrow.” It’s that shipping hardware is the line between “demo” and “ecosystem pressure.” If real volumes follow, it becomes one more force reshaping the global GPU landscape—especially inside China. 

🧠RESEARCH

TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times

Generating AI videos is typically a slow process. TurboDiffusion speeds this up by 100 to 200 times without hurting video quality. It uses faster math for "attention" (how the model focuses), simplifies the generation steps, and reduces the memory size of the model to make it run incredibly fast.

Learning to Reason in 4D: Dynamic Spatial Understanding for Vision Language Models

AI models struggle to track how objects move and interact in 3D space over time. This paper introduces a new dataset and a "Geometry Selection Module" to teach models about dynamic 3D scenes. This allows the AI to reason about movement and changes without getting overwhelmed by unnecessary data.

DreaMontage: Arbitrary Frame-Guided One-Shot Video Generation

Creating "one-shot" videos (continuous, uncut scenes) is hard to fake digitally. DreaMontage allows users to generate these seamless videos using simple prompts or images as guides. It uses smart editing techniques to stitch different segments together smoothly, ensuring the motion looks professional and cinematic throughout the entire clip.

🛠️TOP TOOLS

Each listing includes a hands-on tutorial so you can get started right away, whether you’re a beginner or a pro.

Beducated – AI Sex Coach - online learning platform focused on sexual health, intimacy, and relationships

Beey – AI-Powered Transcription -web-based speech‑to‑text tool for turning audio/video into editable transcripts, captions, and subtitles

Best AI Image Generator Discord- Generate Outstanding AI Images -Image Generator Discord is a web and Discord-based AI platform for creating and editing images.

📲SOCIAL MEDIA

🗞️MORE NEWS

Japan AI Budget Support The Japanese government plans to spend over $6 billion to help local companies build their own artificial intelligence technology. A new group led by major firms like SoftBank will use this money to create powerful AI systems and computer chips within Japan. This move aims to help the country catch up with the United States and China in the tech race.

Data Centers and Jet Engines Tech companies are now using jet engines to generate their own electricity for data centers because local power grids are taking too long to connect them. These facilities need massive amounts of energy to run AI programs, and waiting years for a standard power hookup is no longer an option. This desperate measure highlights the urgent energy crisis facing the booming AI industry.

OpenAI Hiring Head of Preparedness OpenAI is looking to hire a new safety leader to stop its advanced computer programs from causing major disasters like cyberattacks. The job pays over half a million dollars a year but comes with a warning that the work will be extremely stressful. This person will be in charge of testing the technology to ensure it does not help bad actors harm the world.