• Future Tools
  • Posts
  • Our tech future, according to CES đź”®

Our tech future, according to CES đź”®

Plus: Alexa heads to the browser

Welcome back and happy new year! Quick question: When do you think the AI bubble will pop? Late 2026? In a few years? Maybe…never?

I’m teaming up with The AI Collective, Kalshi, and Roam to put that question to the test. If you want to hear where I think AI is heading and weigh in yourself, come join us live for the virtual event next Thursday at 3pm ET!

What’s Standing Out So Far at CES 2026

WIRED / Patrick T. Fallon / Getty Images

Hi from Las Vegas! đź‘‹ CES 2026 is underway, and before the show floor even opened, companies started rolling out a wave of new products, concepts, and AI-powered upgrades. As usual, it’s a mix of genuinely useful ideas, ambitious experiments, and a few things that feel slightly unhinged.

A few AI highlights worth watching:

  • 🤖 LG’s CLOiD home robot: Not just a vacuum, but a mobile, conversational robot with articulated arms designed to fold laundry, fetch items, and act as a roaming smart home hub. It’s less about perfect execution today and more about where embodied AI in the home is heading.

  • đź§Š GE’s AI-assisted smart refrigerator: A fridge that scans empty packages, builds grocery lists, lets you check what’s inside remotely, and integrates cameras and touchscreens to reduce everyday friction. It’s mundane, but very on-theme—AI quietly embedding itself into routine household decisions.

  • đźš— Nvidia’s self-driving car tech: Nvidia unveiled Alpamayo, a new autonomous driving system designed to bring reasoning to vehicles—allowing cars to think through rare edge cases, navigate complex environments, and explain their decisions. 

Why it matters: CES doesn’t usually tell us what will win—but it often hints at where the industry is pushing. This year’s throughline feels clear: AI is moving out of demos and into physical environments, daily routines, and ambient experiences. Less “look what the model can do,” more “how does this show up in real life?”

Amazon Brings Alexa to the Web

Amazon is taking Alexa to the open web. At CES, the company announced the launch of Alexa.com, giving Alexa+ Early Access users a browser-based way to interact with its upgraded AI assistant.

What’s new: This isn’t just Alexa-in-a-tab. The web experience is designed to make Alexa+ more agent-like and central to daily life. Beyond answering questions or generating content, Alexa.com lets users manage calendars, shopping lists, smart home controls, recipes, reservations, and family logistics. 

  • Amazon is also leaning into integrations, adding services like Expedia, Yelp, Square, and Angi alongside existing partners like OpenTable and Uber. 

  • The updated Alexa mobile app mirrors this shift, putting a chat-style interface front and center.

Betting on trust: Amazon’s bigger bet is trust and continuity. Alexa+ encourages users to upload personal documents, emails, and schedules so the assistant can act as a true household hub.

AMD Doubles Down on AI-Powered PCs

AMD used its CES keynote to unveil a new lineup of AI-focused processors, ranging from everyday computing to gaming.

What’s new ⤵️ 

  • The Ryzen AI 400 Series brings 12 CPU cores and 24 threads, with AMD claiming up to 1.3Ă— faster multitasking and 1.7Ă— faster content creation than competitors. 

  • AMD says it now supports more than 250 AI PC platforms, doubling year over year, signaling real momentum behind local, on-device AI workloads. 

  • On the gaming side, the new 9850X3D chip and updated Redstone ray tracing tech aim to deliver better visuals without performance tradeoffs.

The bigger picture: As models get smaller, faster, and more context-aware, companies like AMD are positioning the PC itself as an always-on AI surface. 

Nebius Token Factory—Post-training

Nebius Token Factory just launched Post-training—the missing layer for teams building production-grade AI on open-source models.

  • You can now fine-tune frontier models like DeepSeek V3, GPT-OSS 20B & 120B, and Qwen3 Coder across multi-node GPU clusters with stability up to 131k context.

  • Models become deeply adapted to your domain, your tone, your structure, your workflows.

  • Deployment is one click: dedicated endpoints, SLAs, and zero-retention privacy.

And for launch, fine-tuning GPT-OSS 20B & 120B (Full FT + LoRA FT) is free until January 9. This is the shift from generic base models to custom production engines.

One workspace for every AI model

Via Yonoo

Yonoo is a smart AI router that gives you a single interface for eight frontier models—including ChatGPT, Claude, Gemini, Grok, Perplexity, DeepSeek, and Llama—automatically routing each request to the best model for the task.

How you can use it:

  • Ask one question and get results without switching between models

  • Generate text, images, and video from the same workspace

  • Upload files for research, translation, or analysis

  • Compare outputs across models to choose the best response

Pricing: Free and paid plans available

Build full-stack apps in plain English

Via Caffeine

Caffeine lets you describe an app in natural language, then automatically turns it into a production-ready, full-stack web app. It runs on the Internet Computer (ICP) or private networks, using Motoko and persistent state, so AI-driven updates don’t break or wipe your data.

How you can use it

  • Turn an idea into a live app without writing code

  • Clone and customize starter apps from Caffeine’s App Store

  • Deploy apps with built-in payments, storage, email, video, and search

  • Manage upgrades safely without database migrations or downtime

Pricing: Paid with a free trial

Jobs, announcements, and big ideas

  • Meta delays the international rollout of its Ray-Ban Display smart glasses.

  • Universal Music teams up with Nvidia to build new AI infrastructure for music creation and rights management.

  • LMArena hits a $1.7B valuation just four months after launch and following a $250M raise.

  • xAI secures $20B in Series E funding to scale next-generation AI systems.

  • Lenovo rolls out its Qira AI assistant across laptops and Motorola smartphones.

  • Boston Dynamics partners with Google DeepMind to boost intelligence in humanoid robots.

The Nvidia-Groq “acquisition” explained. I go through what happened and why it matters for the next phase of AI hardware and models.

That’s a wrap! See you Friday for more.

—Matt (FutureTools.io)