Welcome to The Wire. Each month, we’ll deliver fresh insights into the world of learning, health, and work — combining news and data we’re tracking, perspectives from our team, and updates from our portfolio. Subscribe for the latest.

After a decade backing transformative companies from San Francisco, we’ve opened our first East Coast outpost! SoHo will be home for our Partners (Steve Kupfer, Shauntel Garvey, Jomayra Herrera), a gathering spot for local founders, and a window into the vibrant innovations happening across education, healthcare, and the future of work.

New York’s unique collision of entrepreneurial grit, cross-industry pollination, and everyday hustle and bustle creates the kind of creative friction that fuels category-defining companies. We’d love to connect if you’re in the neighborhood!

Congratulations to Replit on its $250 million Series C fundraise led by Prysm Capital at a $3 billion valuation!

We’ve been vibing along since our pre-seed investment in 2016, and we’ve watched the company rocket to $150M ARR while staying true to its mission: empowering the next billion software creators.

“Our vision from the start has been to be a place where you can write your first line of code,” Replit CEO Amjad Masad said in an earlier interview with Reach. “Where you can earn your first dollar, and build your first startup.”

Today, with over 40 million users from creative solopreneurs to enterprises like Anthropic and Zillow, Replit continues to relentlessly innovate — most recently with Agent 3, the first “full self-driving” autonomous software developer that can build, test, and fix apps with 10X cost efficiency.

Personalized tutors. Robo-therapists. These are some of the many “obvious” use cases for AI and LLMs where, as it turns out, building good solutions is anything but simple.

Beneath the surface, these domains are difficult to execute well. They require deep subject expertise and human empathy. Tutoring and therapy aren’t just about answering questions, but building rapport and adapting to different backgrounds. Knowing when to stay silent is just as important as giving advice.

James Kim explores why these problems can’t be fixed through better prompts or data labeling. They require a different approach that combines domain expertise, human services, data sophistication, and cutting-edge AI/ML engineering. For companies that can build this flywheel, the opportunity is enormous: to radically improve how we learn, heal, and grow. 

“People want America to be the best. To stay the best, we have to keep investing in research and keep driving forward to solve every problem that we can.”

Earlier this year, Arizona State University President Michael Crow sat with Wayee Chu to reflect on his path toward building his “New American University,” hailed as one of the most innovative endeavors in higher education. It’s a journey shaped by attending 17 schools before college, an Eagle Scout service project, and a commitment to ensuring that everyone can access the benefits of human and scientific progress.

For Crow, colleges aren’t just centers of learning; they’re bedrocks of innovation, technological breakthroughs, economic advancement, and global competitiveness.

Read the highlights from the conversation here, and watch the video above.

Stories we’re buzzing about in the office.

Fresh data from two of the world’s most widely used chatbots offers a look into how AI tools are reshaping how people learn and work.

Of the two, Anthropic’s latest Economic Index report takes a deeper dive into how Claude is being used across countries and within the U.S. Based on millions of Claude conversations, the accompanying data viz is also chock full of granular insights; a few that caught our eye:

OpenAI’s report has less geographic-specific details, but also reveals some clear — though not surprising — differences in its usage activity. Notably, writing dominates work-related tasks (40% of usage), while coding accounts for just 4%. But it sees a similar uptake for learning: about 10% of all ChatGPT messages are requests related to tutoring or teaching.

If you’re building solutions to move the needle on these figures, or in other areas across learning, health, and work, we’d love to hear from you!

TRADES TRENDING: Vocational degrees among 17-21 year-olds just hit their highest level since 1992 and have doubled since 2000, fueled by TikTok influencers turning welding torches into viral content. Some schools say a staggering 80% of students find them through social media.

MUSICAL SCRUBS: According to a Strategic Education survey, 55% of healthcare workers plan to change roles within a year, citing inadequate pay (49%), burnout (48%), and lack of advancement (48%). Yet while 75% of employers offer education benefits, fewer than 40% of employees use them.

THEORY VS. PRACTICE: A study of 20,000+ students found post-pandemic tutoring delivered only 1-2 months of learning gains versus a full year in pre-pandemic research. The issue: execution. Schools couldn't deliver the high-frequency sessions that make tutoring work, despite the investment in intensive tutoring post-pandemic.

Big moves start with bold clicks. Where will you go next?

There are 708 job openings across the Reach portfolio. Check them out here.