Gutenberg Lab — Morning Paper

Organizing local supporters at permitting hearings scaled clean energy deployment to 67 gigawatts—proving the bottleneck isn't technology

Participants

David Roberts (Volts podcast host) × Matt Traldi (CEO, Greenlight America)

Bottom line

The biggest obstacle to clean energy deployment is not technology or cost, but the simple fact that opponents show up to local permitting hearings while supporters don't—a problem solvable through organizing, not just policy reform.

3-Line Summary

Greenlight America mobilizes local supporters at county-level permitting meetings, advancing 4.4 GW in 2024 to 67 GW cumulative by October 2026, with staff growing from ~20 to nearly 100 and win rate improving from 60% to 79%. The organization now faces an unprecedented sprint: 216-240 GW of solar and ~20 GW of wind have "safe harbored" under expiring tax credits and must complete permitting by early-to-mid 2028 to go into service by end of 2029. This creates more local approval meetings in the next 15 months than any previous period, while data center backlash threatens to contaminate all clean energy with AI industry resentment.

3 Key Points

1. Explosive growth trajectory demonstrates systematic learning

Greenlight advanced 4.4 GW in 2024, 20.9 GW in 2025, totaling 67 GW by October 2026. Staff grew from ~20 to nearly 100, annual funding from ~$5 million to almost $20 million. Win rate improved from 60% (pilot) to 75% (2025) to 79% (2026 YTD). Traldi: "It's really that trajectory over time that shows, okay, we're figuring out how to do this effectively." They went from 4 local wins per quarter in early 2025 to 8 wins in a single day in October 2026. The organization tracks 3,100+ county government websites and works with ~300 local partner organizations across the political spectrum—from faith-based groups to conservative property rights advocates to labor unions.

2. Tax credit cliff creates unprecedented permitting sprint

The 2025 reconciliation bill (OBBBA) phased out wind/solar tax credits but allowed projects starting construction by July 2025 to qualify if they go into service by end of 2029. 216-240 GW of solar plus ~20 GW of wind "safe harbored" by meeting the construction deadline. Working backwards with 12-18 month construction requirement means permitting must complete by early-to-mid 2028. This is "more than a few years of the pace we had been building at before" (~50 GW annually). The next 15 months will see more local approval meetings than any previous period. Industry expects only 20-30% temporary decline post-2029 due to rising power demand.

3. The "feeling in the room" trumps constituent preferences

Traldi observes: "If the first five people speak up in favor of something, it feels like it's sailing through. If the first five people say negative things about it, it feels like it would be political suicide to go forward. And that's true regardless of what decision makers might have felt going into the conversation." This explains why small numbers of organized opponents can block projects despite broader support. A mid-sized utility solar project equals ~4 million trees in emissions impact, and "often at a county council meeting, if you have five people, five local volunteers show up and speak, that makes the difference. So that's like you spend an afternoon, you plant 800,000 trees."

Editorial Perspective

Local politics remains surprisingly non-partisan despite national toxicity—even conservative decision makers in conservative areas decide based on "is this good for my community" rather than national partisan talking points. This suggests local deployment can proceed even in hostile national political environments. However, data center backlash is creating dangerous spillover: Traldi reports "many potential clean energy supporters, natural allies, if you will, are either flipping to opposition or staying on the sidelines because of a perceived connection between clean energy and data centers." The specific misunderstanding—"if a solar project gets built, then that will encourage the data center"—threatens to contaminate all electrification with AI/tech industry resentment. This spillover may be a more serious threat than any organized opposition.

Source: Volts, "Checking back in with Greenlight America" (October 14, 2026)

https://www.volts.wtf/p/checking-back-in-with-greenlight

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

Can Open Models Keep Up With the AI Frontier? — The Structural Gap in Capital and Execution

Featuring

Lisha Li (a16z Partner) × Nathan Lambert (Founder of Trillium, former post-training lead at AI2)

Bottom line

The gap between open models and frontier labs has narrowed to 2-3 months, but structural disparities in capital, compute, and organizational execution persist.

3-Line Summary

The lag between open-source AI models and frontier labs (OpenAI, Anthropic) has shortened to 2-3 months. Chinese companies match US labs in technical talent but face orders-of-magnitude compute shortages—an average OpenAI researcher's access to ~5,000 H100 equivalents made a top Chinese pre-training researcher's "jaw fall to the floor," Lambert notes. Distillation is less effective than conventionally believed for catching up, while frontier labs maintain revenue through integrated stacks and first-mover advantages. The open model economy is fragmented and slower but offers model ownership and verifiable ROI.

3 Key Points

1. Compute disparity is the defining constraint: Chinese labs match US frontier labs in technical talent but face severe compute shortages. An average OpenAI researcher has access to ~5,000 H100 equivalents—a figure that made a top Chinese pre-training researcher's "jaw fall to the floor." This compute gap, not talent or technique, is the primary barrier preventing Chinese labs from reaching the frontier.

2. Distillation's effectiveness is overstated: Lambert challenges the conventional wisdom that distillation is the main lever for open models to catch up. Based on open literature, researchers are "stuck using outdated Qwen models" because they can't get the teacher model right. It's "hard to map" traditional distillation (supervised fine-tuning on completions) to current reasoning model training. He's "not sure" that access to Claude's reasoning traces would have significantly improved his AI2 models.

3. The data economy has a built-in lag mechanism: Data companies sell to OpenAI and Anthropic first at premium prices, then sell similar datasets to others after contractual periods (often ~1 year) expire. This creates a structural 6-12 month advantage for frontier labs but ensures catch-up is "very doable." The fast-following cycle has lasted longer than expected—Lambert thought consolidation would happen two years ago.

Editorial Perspective

The frontier lab advantage lies not in capital quantity but in the ability to organize capital and execute. The fact that Chinese companies match technical talent but lag orders of magnitude in compute reveals that AI competition is a resource allocation contest, not purely technical. Meanwhile, Silicon Valley's information security is loose enough that "ideas travel like one beer"—frontier techniques diffuse within months. Open models take longer but offer ownership and verifiable ROI; frontier labs secure revenue through integrated stacks and first-mover advantages. Both operate on different exponential timescales, and neither will disappear.

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

Source

This Week in Startups, "Can Open Models Keep Up With the AI Frontier? | Nathan Lambert" (October 14, 2026)

https://a16z.simplecast.com/episodes/can-open-models-keep-up-with-the-ai-frontier-nathan-lambert-WqIVvxfv

TV's Most Successful Sitcom Creator Explains Why the Format Died—and Why Success Can't Fill the Hole Inside

Featuring

Joe Rogan × Chuck Lorre (creator of The Big Bang Theory, Two and a Half Men, Young Sheldon)

Bottom line

External success—fame, money, adulation—cannot heal internal wounds, and when damaged people discover this truth, they face only two paths: drugs or a spiritual practice.

Three-Line Summary

• Who: Chuck Lorre, creator of The Big Bang Theory and Two and a Half Men, discussed the collapse of the multicam sitcom format and the toxic dynamics behind even massively successful shows
• What: The decline from 40-50 multicam sitcoms per week in the 1990s to 2-3 today is not a cultural shift but structural mismanagement by television networks
• Why it matters: The mechanism by which successful actors and comedians fall into addiction (the despair of realizing success doesn't fill the void) and the spiritual path as the only solution reveals how to survive Hollywood's toxicity

Three Key Points

1. The sitcom extinction event: In the 1990s, four networks aired four sitcoms each, five nights a week—40 to 50 multicam sitcoms weekly. Today, "maybe three... maybe two" remain. Lorre believes "one great show changes everything," but Rogan counters it's "mismanagement of television" that allowed sitcoms to vanish despite global demand (The Big Bang Theory remains a massive hit in Germany, Mexico, and Brazil).

2. The laugh track myth as delegitimization: Lorre has been accused of using laugh tracks for forty years but has never used one. A reporter sat in The Big Bang Theory's live studio audience, watched 200 people laugh for 3.5 hours, then wrote an article complaining about the laugh track. Lorre believes the word "sitcom" itself is weaponized to diminish the format compared to prestige single-camera shows.

3. The moment success kills: "If you're damaged in this certain kind of way that actors and comics can be damaged and you succeed and you learn that horrible, horrible 2 o'clock in the morning horrible truth, I'm still unhappy. Success was supposed to solve my problems... Now what? If that's not gonna solve that hole inside." Lorre traced this pattern through Brett Butler (Grace Under Fire), Charlie Sheen, and his own addiction, concluding: "Nothing on the outside is gonna heal the inside. It has to be an inside job."

Editorial Perspective

Lorre's principle that "the outside can't heal the inside" is a pattern we see repeatedly in startup culture during platform transitions. Funding rounds, unicorn valuations, media coverage—all temporarily fill the void but never answer the fundamental question: "What are we building this for?" Founders who chase success without that anchor find emptiness waiting at the finish line. Lorre spent ten years in Scientology and another 27.5 years sober to reach his definition of spirituality as "a verb, not a noun"—a refusal to seek answers externally. The death of the sitcom may also stem from trying to measure laughter (a verb) with ratings (a noun). When the format that once produced 40 shows weekly now produces two, the question isn't whether audiences still want to laugh—Lorre's global success proves they do—but whether the industry still knows how to let them.

Source

The Joe Rogan Experience "Chuck Lorre" Episode 2564 (October 13, 2026)

(unknown)

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --