😺 Meta Update: The Metaverse is dead. Long Live Superintelligence...

Welcome, humans.

Wharton’s Ethan Mollick showcased generative UI in practice via a custom Claude Code plugin. It visualizes backend tasks as office workers who are “hired,” gain skills, and collaborate—Sims meets The Office.

As we predicted, This is a version of what we predicted about generative UI becoming a huge deal this year. Generative UI is about creating the exact interface you need on demand. Mollick didn’t wait for Anthropic; he simply prompted Claude to build a custom “under the hood” view.

Soon, every tool will be this flexible. Imagine turning your inbox into a Kanban board or workflows into an RTS game via simple prompts. The interface becomes as adaptable as the task, and its creation trivial.

Here’s what happened in AI today:

• We review everything going on at Meta atm.

• Big Tech’s 1,500+ energy hires.

• Google’s MedGemma-1.5 for health data.

• Anthropic launches Labs.

ICYMI: We just interviewed IBM’s David Cox on their new Granite 4.0 open models. We covered their hybrid architecture, ISO certification (a first for frontier models), their 10x memory efficiency, the ability to swap fine-tune “adapters” in and out, and why labs should stop chasing AGI. “Why do we even want that?”

Watch and/or Listen now: YouTube | Spotify | Apple Podcasts

What’s up with Meta? The Death of the Mertaverse, Feuding Scientists, and a new Deal Under Investigation

The metaverse is on life support, sacrificed for an infant superintelligence. CNBC reports Meta’s “major course correction”: 1,000 VR jobs cut and studios shuttered to fuel a $70B AI pivot.

Time of Death? Between Sanzaru Games closing and the latest $14B AI wire transfer. It’s a Silicon Valley ghost story: sacrificing an imaginary world to summon an imaginary god. What’s going on at Meta?

The Llama 4 Confession: Ex-chief AI scientist Yann LeCun torched Meta on his way out. In a recent interview, he claimed Meta “fudged” Llama 4 benchmarks by mixing model versions for better results. Mark Zuckerberg reportedly “lost confidence” in the GenAI org, leading to mass departures. LeCun also slammed his replacement, 28-year-old Alexandr Wang (Scale AI), as “young” and “inexperienced.” His parting shot: “You don’t tell a researcher like me what to do.”

The Manus Investigation: Meta’s $2B Manus acquisition is under fire from Chinese regulators. Manus “Singapore-washed” (moved from China to Singapore) last year, and China’s Ministry of Commerce is investigating potential export control violations. If found guilty, the founders face criminal charges and the deal could collapse.

What’s Shipping: Meta isn’t all chaos. SAM 3D now powers Facebook Marketplace’s “View in Room” feature, turning single photos into 3D. Two 2026 flagship models—Mango (media) and Avocado (reasoning)—are coming via Wang’s Superintelligence Labs. Additionally, “Meta Compute” (led by Santosh Janardhan and Daniel Gross) aims for tens of gigawatts of AI infrastructure this decade.

Our Take: Meta is a paradox: massive capital vs. departing talent; bold acquisitions vs. regulatory heat. With high stakes and past flops, all eyes are on Alexandr Wang to deliver Zuck’s superintelligence. We’ll find out soon.

FROM OUR PARTNERS

Vellum: AI agents for your boring ops tasks

Last year AI Agents finally started working for engineers. Cursor, Devin, Claude Code gave them real leverage. Meanwhile, the rest of us are stuck in the same loop since 2023: lots of ChatGPT tabs, lots of context switching, not much has changed. 

That changes today.

Introducing Vellum 🎉

Describe the task you want to automate and Vellum turns it into a working agent (like Lovable for agents).

Vellum will ask follow-up questions to understand what you need and connect to your tools. It handles the logic and makes the agent’s behavior visible, so you’re never guessing what it’s doing or why.

Vellum is launching today and we’re excited to share this with the Neuron community. To celebrate, everyone who signups today will get 30 extra credits (worth $100) to try it out.

Get early access here

Prompt Tip of the Day

Claude Code now supports custom instructions via the Tab key on permission prompts (e.g., “yes, if it passes X tests”). This ends the binary choice between “YOLO-ing” agents and tedious babysitting. Also, read Dan Shipper’s Agent-Native Architectures or catch our TL;DR actionable version in the Prompt Tip of the Day Digest.

FROM OUR PARTNERS:

Editor’s Pick: SurveyMonkey

SurveyMonkey helps you create surveys in seconds and turn responses into clear insights, so you can collect customer, employee, or market feedback and instantly spot themes, sentiment, and trends without manual analysis.  

Join their 30-minute webinar | January 14 | 10am PT/ 1pm ET here.

Treats to Try

1. GLM-Image generates images with accurate text and knowledge-dense details—tops benchmarks for rendering multiple text regions correctly (blog, API, HuggingFace). Free to try.

2. Willow lets you dictate anywhere (even offline) and auto-learns names/terms. Free trial, then $15/mo (raised $4.5M).

3. NEAR Private Chat runs OpenAI/DeepSeek models in hardware-sealed TEEs for encrypted conversations (blog).

   • NEAR AI Cloud provides cryptographically proven secure inference for 100M+ users. Pricing: DeepSeek V3.1 ($1/M tokens), Qwen3 30B ($0.15/M).

4. Confer from Moxie Marlinspike encrypts chats via device-only keys in TEEs; verifiable open-source builds (blog).

   • Difference: NEAR targets enterprise; Confer is consumer-focused/fully open-source.

5. Pindrop detects deepfake voice/video in real-time with 99% accuracy.

6. Eigent open-sourced Open Cowork, which splits workflows across specialized agents. Also: SETA provides 400 environments for terminal agents. Free.

7. MedGemma 1.5 from Google analyzes clinical data/images to mark abnormalities (blog). Free.

8. Atoms builds complete businesses autonomously—from multiplayer games to home design sites.

Around the Horn

1. Apple and Qualcomm are in a race to secure Japan’s glass cloth supplies amid shortages lasting through 2027 due to AI chip demand.

2. US Defense Secretary Pete Hegseth announced Grok’s deployment on military networks this month alongside Gemini (despite deepfake controversies).

3. Samsung is reviving Bixby with Perplexity AI for complex queries in One UI 8.5 beta.

4. Cerebras plans to raised $1B at a $22B valuation ahead of its IPO.

5. Big Tech hired 1,500+ energy experts since 2022 to manage AI infrastructure demands.

   1. Related: US President Trump warned tech firms must pay for data center power; Microsoft pledged to cover infrastructure costs.

6. Anthropic launched Labs under Mike Krieger to accelerate frontier product shipping. Mr. Kill Your Startup is hittin’ the town…

FROM OUR PARTNERS

Future Proof with AI - Harvard's AI Learning Platform

Build practical AI skills—fast. The Digital Data Design Institute at Harvard has created Future Proof with AI for professionals navigating an AI-powered world. Learn on-demand, earn completion badges, and apply AI to real business decisions. Full platform access until September 2026 for only $199 with code NEURON.

Future-proof your career with AI: Enroll now for $199 using code NEURON

Midweek Meme

If you’re looking for Midweek Wisdom, check out our new Research Digest for January here.

A Cat’s Commentary