Good morning. It's Tuesday, March 31, and we're covering robots scaling solar construction, AI-driven layoffs reshaping tech jobs, Claude stepping closer to full computer control, and more.

If you enjoy this email, share it with a friend. First time reading? Sign up here to keep up with the future of tech.

Claude CLI Gets Full Computer Control
Claude Code now lets the AI control macOS apps via the CLI in a research preview. It can open apps, click through interfaces, and debug GUI workflows without APIs. Users must approve apps per session for safety. The feature prioritizes other tools first, using screen control only when needed. It marks a step toward AI handling full end-to-end software tasks.

Suno’s AI Music Growth Explodes
AI music startup Suno has reached $300M ARR with around 2 million paying users. The platform is generating roughly 7 million songs per day, highlighting rapid adoption. At that pace, output rivals Spotify’s full catalog every two weeks. The figures suggest explosive growth in generative music, though claims are unverified. It underscores rising competition between AI-generated and traditional music ecosystems.

Rebellions Raises $400M Ahead Of IPO
South Korean AI chip startup Rebellions raised $400 million at a $2.34 billion valuation, backed by investors including Samsung-linked funds. The company is targeting U.S. expansion and preparing for an IPO, focusing on AI inference chips that compete with NVIDIA. It is already running trials with major AI labs but faces memory supply constraints. The raise signals surging demand for energy-efficient AI infrastructure

Mistral Raises $830M For Data Center
Mistral AI has raised $830M in debt to build a data center near Paris, expected to launch in Q2 2026. The facility will run on NVIDIA chips and support growing demand for AI compute. It’s part of a broader plan to deploy 200MW of capacity across Europe by 2027. The move marks Europe’s push toward AI independence from U.S. cloud providers.

Matt explores 4 viral open-source AI agent projects—GStack, Hermes, Superpowers, and Paperclip—reshaping coding, workflows, and autonomous companies.

At AES’s Bellefield solar complex in California’s Mojave Desert, a fleet of Maximo robots has installed 100 megawatts of solar capacity—one of the largest real-world deployments of field robotics in energy construction. Using Version 3.0 systems, human crews assisted by robots placed up to 24 photovoltaic modules per hour per worker, with robots assembling more than one panel per minute—nearly doubling typical installation speeds in the region.

The project shows how automation is moving beyond pilot stages into utility-scale infrastructure. It also reflects broader pressures on the energy sector, including labor shortages and rising demand for clean power driven by data centers and electrification. → Read the full article here.

Major tech companies including Amazon, Meta, and Block are increasingly attributing job cuts to advances in artificial intelligence, reframing layoffs as a byproduct of productivity gains rather than cost-cutting. Meta alone has cut hundreds of roles in recent weeks, with CEO Mark Zuckerberg predicting 2026 will significantly reshape work through AI.

Executives argue smaller teams can achieve more using AI tools, with some firms reporting 25%–75% of code now AI-generated. However, analysts note AI is also serving as a convenient narrative as companies ramp up massive spending—projected at $650 billion—on AI infrastructure. The shift signals both real efficiency gains and a strategic repositioning of layoffs amid investor pressure. → Read the full article here.

Chinese tech giants including Alibaba, Tencent, and ByteDance are pouring billions into promotions to drive everyday use of AI chatbots, prioritizing adoption over cutting-edge performance. During the February 2026 Lunar New Year, companies collectively spent over $1.1 billion on incentives—like free milk tea—to attract users and normalize chatbot-based transactions.

Apps such as Alibaba’s Qwen and ByteDance’s Doubao integrate payments, shopping, and services directly into chat interfaces, aiming to become “everything apps.” The strategy has driven massive spikes in usage—Qwen hit 73.5 million daily users, while Doubao surpassed 144 million—but retaining users remains a challenge after promotions end. The push reflects China’s broader strategy of embedding AI into daily digital life at scale. → Read the full article here.

Researchers from Stanford, MIT, and collaborators introduced Meta-Harness, a system that automatically optimizes the “harness” surrounding large language models—the code that governs how models store, retrieve, and process information. The approach uses an agentic loop to iteratively rewrite and test harness code, leveraging full execution histories rather than compressed feedback.

In experiments, Meta-Harness improved text classification accuracy by 7.7 points while using 4× fewer context tokens and boosted performance on advanced math and coding benchmarks. The system also outperformed hand-designed and prior automated methods on TerminalBench-2. The results suggest that optimizing the surrounding system, and not just the model, can yield significant gains. → Read the full paper here.

Qwen3.5-Omni Expands Multimodal AI: Alibaba unveils a model handling text, image, audio, and video with 100M+ training hours and real-time interaction tools, intensifying competition in full-spectrum AI systems.

DeepSeek Outage Disrupts Millions: Chinese AI startup’s chatbot went offline overnight on March 30, 2026, affecting 355 million users before restoration, raising reliability concerns as rivals gain ground.

Qodo Raises $70M for Code Verification: New York startup secures Series B to build AI agents ensuring reliability of AI-generated software, as enterprises confront trust and governance gaps in automated coding.

Delaware Judge Reassigns Musk Cases: Chancellor Kathaleen McCormick moves three Tesla-related lawsuits, citing media bias concerns after LinkedIn controversy, spotlighting scrutiny over judicial impartiality.

Mantis Biotech Builds Human Digital Twins: New York startup raises $7.4M to create physics-based synthetic human models from fragmented data, aiming to unlock research in rare diseases and data-scarce medical scenarios.

Thanks for reading. See you next time!

— Matthew Berman, Nick Wentz & the Forward Future Team