Zuck says AI will remake work in 2026

Here’s what’s worth your time today.

[2 min read]

Mark Zuckerberg says 2026 will be the year AI fundamentally changes how work gets done at Meta. As the company flattens teams and pours money into AI-native tools, Zuckerberg argues that individual employees can now do what once required entire groups. Meta says output per engineer is already up 30% since early 2025, with heavy AI users seeing gains as high as 80%. The bet is that AI-driven productivity will help justify massive spending on data centers and chips while making Meta a more attractive place for top talent. Behind the optimism is pressure: Meta expects capital expenditures to climb as high as $135B in 2026, and AI needs to pay off both internally and in the products that power its ad business.

Supported by AssemblyAI

Accuracy, conversational understanding, and low-latency responses consistently rank as the top priorities for teams building voice agents. But many solutions force tradeoffs between performance, integration flexibility, and cost.

The Voice Agent Report highlights what matters most:

AssemblyAI supports production-grade voice agents designed for real-time conversations, without unnecessary complexity.

[2 min read]

Anthropic is expanding its partnership with ServiceNow, with Claude models now powering ServiceNow’s Build Agent for creating apps and workflows using natural language. The deeper integration positions Claude as a trusted AI layer inside enterprise systems, operating within existing security, compliance, and monitoring rules. ServiceNow says Claude’s coding strength and domain expertise will help it serve regulated industries like healthcare and life sciences, while keeping model choice open for customers. The deal highlights a broader shift: AI that’s embedded directly into core workflows is becoming a revenue driver, not just a feature, as enterprise buyers look for tools that actually save time and money.

[3 min read]

OpenAI has unveiled Prism, a new AI-powered workspace designed to help scientists write, review, and refine research papers. Available free to anyone with a ChatGPT account, Prism works like an AI-enhanced word processor, tightly integrated with GPT-5.2 to assess claims, revise prose, search prior literature, and manage full project context. OpenAI positions it as the science equivalent of coding tools like Cursor or Windsurf. The bet is that tighter workflow integration, not just better models, will speed up human-led research. As scientific queries surge inside ChatGPT, Prism signals OpenAI’s push to make AI a default layer in how research gets done.

[2 min read]

Microsoft reported another blowout quarter, with revenue up 17% to $81.3B and profits jumping 60%, even as it spent a massive $37.5B on AI-related data centers in just three months. Azure grew 39%, reinforcing how central AI has become to Microsoft’s business, but shares still fell more than 5% after hours as investors focused on costs and ongoing capacity constraints. CEO Satya Nadella said AI demand continues to outstrip supply and warned those limits will last through 2026, underscoring how expensive it is to stay at the front of the AI race, even when the numbers look strong.

[1 min read]

Tesla is winding down parts of its EV business as it shifts focus toward robotics and artificial intelligence. The company reported its first annual revenue decline, down 3% in 2025, and said profits fell sharply at the end of the year. Tesla will end production of the Model S and Model X and repurpose its California plant to build its humanoid robot, Optimus. At the same time, Tesla disclosed a $2B investment in xAI, deepening its bet on AI even as shareholders remain divided. With BYD overtaking Tesla in global EV sales and political backlash weighing on the brand, the move signals a clear reset. Tesla is betting its future less on cars and more on robots, AI, and capital heavy moonshots.