The AI optimists vs. skeptics ⚖️

Plus: Creators go all-in on AI clones

January 23, 2026

Welcome back! CEOs say AI will change everything. But employees? Not so much. In a new survey of 5,000 white-collar workers, 40% of non-management employees report that AI saves them zero time each week compared to 19% of executives.

It brings up a big, gnarly question: Are we building AI for the people who buy it, or for the people who use it? Hit reply and tell me what you think.

Tech Leaders Diverge on AI at Davos

Image Credits: Benjamin Girette / Bloomberg / Getty Image

The World Economic Forum in Davos became a parade of AI predictions this week, with tech leaders from Nvidia, Microsoft, Anthropic, and Google DeepMind taking the stage to share their (very) different visions of what's coming. Here are some of their predictions. ⤵️

The optimists: 

  • Nvidia's Jensen Huang declared that AI has the “largest infrastructure buildout in human history” and is creating a boom in manual labor jobs—plumbers, electricians, construction workers—with salaries nearly doubling to six figures in some cases. 

  • Google DeepMind's Demis Hassabis predicted new, more meaningful jobs being created, though he acknowledged we'd be in “uncharted territory” once artificial general intelligence arrives in 5–10 years.

The cautious: 

  • Anthropic CEO Dario Amodei warned that the next few years will be critical for regulation and governance, saying AI is “knocking on the door of incredible capabilities." He told Bloomberg that half of all entry-level white-collar jobs could be wiped out and argued that "not selling chips to China is one of the biggest things we can do" to buy time. 

  • Philosopher Yuval Harari went further, warning that AI will never be like humans and that "the most intelligent entities on the planet can also be the most deluded."

The bigger picture: Tech leaders can't agree on timelines, job impacts, or what to do next. The only consensus? We're building faster than we understand, and the consequences won't wait for us to figure it out.

YouTube Will Let Creators Clone Themselves With AI

YouTube is about to flood Shorts with AI deepfakes. But this time, creators are making them. CEO Neal Mohan announced that creators will soon be able to generate Shorts using their own AI likenesses.

The AI creator toolkit expands: This is the latest in YouTube's aggressive push into generative AI for creators. The platform has been steadily rolling out tools such as auto-dubbing, AI stickers, and AI-generated clips to make content production faster and more accessible. 

But YouTube is also cracking down: The irony is hard to miss. While YouTube encourages creators to clone themselves, it's simultaneously fighting unauthorized AI clones. Last October, the platform rolled out likeness-detection technology that identifies AI-generated content featuring a creator's face or voice without permission.

Why it matters: Platforms are trying to thread an impossible needle: embrace AI-generated content to stay competitive with TikTok and Instagram, but prevent the flood of low-quality AI slop that degrades the user experience. Once you give creators the tools to mass-produce AI versions of themselves, you've opened the floodgates.

Adobe Turns Acrobat Into an AI Command Center

Adobe just announced new AI-based features that let users edit PDFs using text prompts, generate podcast summaries of documents, and create entire presentations from files stored in Adobe Spaces, its collaborative file storage system launched last year.

How it works: If you have financial details, product plans, and competitor analysis saved in an Adobe Space, you can prompt Acrobat's AI assistant to build a pitch deck that positions your product against rivals. The AI generates an editable presentation outline, which you can then style using Adobe Express' theme library, stock photos, or your own images. 

Why it matters: Adobe is changing how people interact with documents. Instead of manually editing PDFs or building presentations slide by slide, users can now describe what they want and let AI execute. 

Web agents on autopilot

Via Gobii

Gobii deploys AI “digital workers” that browse the web 24/7, run JavaScript, extract and synthesize data, and trigger actions on schedules or events (even on sites without APIs). 

How you can use it:

  • Monitor competitors, pricing pages, or regulatory sites continuously

  • Automate lead research and enrichment for sales and recruiting

  • Run compliance and risk checks on a recurring cadence

  • Feed structured web data into internal tools and dashboards

Pricing: Free and paid plans

Turn docs into data

Via Parserdata

Parserdata uses AI to extract structured data from invoices, receipts, and bank statements, including scanned PDFs and images. It then exports the data into clean Excel, JSON, or XML. 

How you can use it:

  • Automate accounts payable and expense processing

  • Generate financial summaries and anomaly signals

  • Build structured datasets from document archives

Pricing: Free and paid plans

Jobs, announcements, and big ideas

  • Congress goes after ChatGPT’s marketing claims as AI ads enter the spotlight.

  • Waymo turns Miami into its next real-world testbed for fully driverless rides.

  • Google quietly absorbs Hume’s emotion AI team to push Gemini beyond text and images.

  • DeepMind’s CEO draws a line on ads as OpenAI tests monetizing the chatbot layer.

  • OpenAI moves to claim a cut of what customers build on top of its models.

  • Spotify’s AI playlists roll out globally, letting prompts shape what you hear next.

Why I ignored the robots at CES: I break down my findings at CES 2026, from AR wearables and robotaxis to humanoid robots.

That’s a wrap! See you next week.

—Matt (FutureTools.io)