12 days of Shipmas + a new Czar
Are we shipping AI features faster than anyone can use them?
AI shipping speeds are blowing past user adoption rates - and that's a problem.
I was reading Cohere CEO Aidan Gomez's shareholder letter today and this jumped out at me: "While consumers have fallen in love with the technology and use it as part of their daily lives, enterprises are struggling to keep up."
For all its hype — and the dizzying pace of new AI products hitting the market — something isn't clicking.
The numbers tell a sobering story. BCG just dropped a report showing that while CEOs are throwing money and talent at AI, only 22% of companies have moved past the proof-of-concept stage. Even wilder? Only 4% are creating real value.
It's making me wonder - are we shipping features faster than anyone can actually use them? It's like we're building Ferrari engines when people are still learning to drive.
At what point does innovation outrun utility?
I see this playing out in two ways:
First, AI companies might finally hit pause on the feature frenzy and zero in on genuine user needs. Imagine that - technology that solves real problems instead of chasing capabilities.
Or (and this is where it gets interesting), we swing so far the other way that choice becomes paralyzing. Your personal AI becomes your gateway to everything, collapsing a thousand apps into one interface that just gets you.
The gap between shipping speed and user value is widening. The question isn't just about catching up - it's about whether we're even running in the right direction.
Product folks: what do you think? Where does this go from here?
The Latest This Week
As per the op-ed, it seems like everyone’s rushing to get their latest announcements out before the holidays!
Infrastructure
OpenAI’s 12 days of ‘shipmas’ include Sora and new reasoning model: OpenAI plans to kick off a “shipmas” period of new features, products, and demos for 12 days, starting on December 5th. The announcements will include OpenAI’s long-awaited text-to-video AI tool Sora and a new reasoning model.
Google launches Veo and Imagen 3 generative AI models: Google Cloud has launched two generative AI models on its Vertex AI platform, Veo and Imagen 3. Veo, available in private preview on Vertex AI, represents a milestone as Google becomes the first hyperscaler to offer an image-to-video model.
Amazon announces its own set of Nova AI models: Amazon announced a series of new AI foundation models under a new “Nova” branding that will be available as part of the Amazon Bedrock model library in AWS. There are three “understanding” models available now: Amazon Nova Micro: a text model that’s “optimized for speed and cost.” Amazon Nova Lite: a “very low-cost” multimodal model that can take in images, video, and text to generate text. Amazon Nova Pro: a “highly capable” multimodal model. The company is also training a model called Amazon Nova Premier, which it says will be “our most capable multimodal model for complex reasoning tasks.”
Gaming & 3D
Fei-Fei Li’s World Labs generate 3D environments from a single picture: World Labs announced a new AI feature that turns 2D images into tiny worlds. This tool, the first from Fei-Fei Li’s highly hyped startup, takes a prompt or 2D reference image and spits out a fully navigable 3D environment. It creates scenes with persistent visual aspects, real-time control, and consistent geometry, allowing creators to create precise camera movements within the generated environment.
DeepMind’s Genie 2 can generate interactive worlds that look like video games: DeepMind’s Google’s AI research org, has unveiled a model that can generate an “endless” variety of playable 3D worlds. Called Genie 2, the model Trained on videos, the model’s able to simulate object interactions, animations, lighting, physics, reflections, and the behavior of “NPCs.”
Agents & Experiences
Meet Runner H: the most advanced AI agent for real-world applications. It doesn’t just outperform competitors in speed and accuracy—it also handles a far wider range of tasks, tackling challenges others can’t.
Humane wants to put the AI Pin’s software inside your phone, car, and smart speaker: Humane, which makes the not-great AI Pin, wants other companies to build AI devices and gadgets that use its CosmOS operating system, and it has released a video that appears to show that the company already has it working in a car, TV, smart speaker, and phone. The CosmOS SDK isn’t available publicly yet – the company’s website only says that it’s “coming soon.
The company behind Arc is building a new AI web browser called Dia: The Browser Company CEO Josh Miller teased in October that it was launching a more AI-centric product, which a new video reveals is Dia, a web browser built to simplify everyday internet tasks using AI tools. It’s set to launch in early 2025.
AI Video
Tencent dropped HunyuanVideo, a new 13B open-source text-to-video model that looks like it is going to take the throne in AI video. The model delivers cinematic video quality in both real and virtual styles. It is breaking the norm with dynamic motion, capturing entire actions in a single shot
Adobe has revealed MultiFoley, an AI system that generates synchronized post-production sound effects for videos through various input methods. It achieves 0.8-second accuracy while maintaining high-quality 48kHz audio output.
VideoDB announces Director, a video agent framework that can help users edit, search for, and even translate videos all through a chat interface.