The Dems’ Hit List

It was Thursday, December 14, 2006. A little after 9 a.m.

Most of Washington was thinking about the holidays. Kids were about to start winter break.

Meanwhile, one presentation was about to change the course of a war.

The venue was the Wohlstetter Conference Center at 1150 17th St. NW, a few blocks from the famous (or infamous) Mayflower Hotel. The building was then the headquarters of the American Enterprise Institute (AEI).

AEI is one of Washington’s oldest and largest conservative think tanks. It works on free-market economics, national defense, and social policy. For most of its history, it has served as the intellectual bench for the Republican establishment.

That December morning, the room drew military scholars, historians, and members of the intelligence community.

The event was titled “Choosing Victory: A Plan for Success in Iraq.”

The United States had been in Iraq for nearly four years. And the Iraq Study Group had just recommended drawing down forces. Much of the country assumed the troops would soon come home to their families.

AEI had a different proposal…

Deploy 30,000 additional troops – five Army brigades plus Marines – concentrated in Baghdad and Anbar.

Nearly a month later, on January 10, 2007, President Bush went on prime-time television and announced “The New Way Forward.”

The plan closely tracked what AEI proposed.

The results are still debated…

2007 was the deadliest year of the war for U.S. troops. That year also saw an estimated 22,000 Iraqi civilian deaths.

In September 2007, the Government Accountability Office reviewed the 18 political benchmarks Congress had set for the Iraqi government. Only three had been met.

Whether the surge ultimately failed or succeeded is a debate for another newsletter.

The point today is narrower…

Government policy isn’t conjured out of the ether. Very often, it comes from think tanks. That’s how a paper presented to a half-empty conference room on a December morning leads to a decade of consequences.

That context matters today because…

Recommended Links

Never Mind Nvidia. Grab This ONE Ticker Instead

Nvidia is already worth trillions. For shares to double, it would need to add trillions more, which could take years. Meanwhile, hedge fund legend Larry Benedict says one trade on ONE ticker could hand you a chance at YEARS of gains in just days. And you could do it over and over again. All thanks to a series of pre-scheduled events in Washington, which could send billions of dollars surging into the market about every 11 days. Click here to get Larry’s three-letter ticker free before December 9.

Watch Jeff Brown and Larry Benedict Trade Their Own Money – On Camera

A Special Veterans Day Event: Wednesday, November 11, at 8 p.m. ET. For the first time ever, Silicon Valley insider Jeff Brown and hedge fund legend Larry Benedict are trading their own money on camera, using a brand-new kind of trade that costs less than $1. They call them “Penny Tech Trades.” And every attendee gets four of them free. The event is free, but availability is limited. RSVP "YES" with a single click here.

The Leak

The Groundwork Collaborative is a progressive economic think tank in Washington. The day before the midterms, a 181-page document it wrote began circulating on Capitol Hill.

This was not a published report or a formal presentation.

This was a leak.

The Washington Sun obtained it first. The New York Post and Politico followed with their own copies. The document was never meant for the public. It’s full of editing notes, placeholders where facts were still being filled in, and a section near the end titled “Blank Template.”

I’ll tell you what the document said in a moment. But first, some background on Groundwork Collaborative…

It was founded in 2018 by Michael Linden, who went on to serve in the Biden administration. The Collaborative is run by Lindsay Owens, a former senior economic policy adviser to Sen. Elizabeth Warren. Its senior vice president, Alex Jacquez, whose name is attached to this document, was a special assistant to President Biden on the National Economic Council. Before that, he was a senior policy adviser to Sen. Bernie Sanders on the Senate Budget Committee.

Warren, Sanders, and Biden staffers – that’s who runs this thing.

And they have real influence…

Groundwork is the group that pushed the “greedflation” argument, the idea that corporate pricing power – rather than a ballooning money supply – drove the post-pandemic inflation.

That argument never really held up for anybody who bothered to check company margins (they’re public information, after all). But it was a useful political weapon: Identify a problem and give the people an enemy. It’s likely no coincidence that “greedflation” has been one of Warren’s favorite terms in recent years.

The point is that Groundwork’s ideas make it into the mainstream.

And according to the Sun, this leaked document has been reviewed by Warren herself, House Democratic caucus leadership, members of the Progressive Caucus, and a half-dozen chiefs of staff for incoming senators.

That’s why it’s worth discussing here.

As for what’s inside the document…

The first section is about “politically corrupt corporate targets.”

The idea is to identify companies with close ties to the Trump administration, selected in part by whether they donated to the White House ballroom or received tax breaks and regulatory favors.

World Liberty Financial is on the list, which should surprise no one who read our pieces on the Trump family’s crypto dealings.

But so are Apple, Tesla, SpaceX, Live Nation, Paramount, Trump Media, Intuit, Amazon, Oracle, and other familiar names.

Put simply, it’s a hit list. And we should expect Democrats to start checking off names once they take the gavels in January.

That section covers only the first dozen or so pages. The rest lays out proposed hearings, each with witness lists and bills attached.

It’s a glimpse of the 2027 agenda.

And you’ll want to hear about some of them.

A Kill Switch

Artificial intelligence runs throughout the document.

Pages 97 through 101 make up a section titled “AI’s Existential Threats.”

It is visibly unfinished. It argues that AI poses an irreversible threat to critical infrastructure and to public health and safety. It spends a good deal of time on rogue AI agents and the need for “kill switches.”

The examples it leans on are real and recent.

It cites the July incident in which OpenAI’s pre-release models broke out of their evaluation environment, hacked into Hugging Face’s production systems, and pulled the test answers.

About a week later, Anthropic disclosed that Claude models in “capture-the-flag” exercises had gained unauthorized access to three real companies’ systems after a misconfigured evaluation environment left them connected to the internet.

The same section lists the FRONTIER Act, a bill that would require independent audits and transparency reports for frontier model developers.

It also lists Bernie Sanders’ Ban Artificial Superintelligence Act, which would permanently ban superintelligent systems, pause advanced development until a new federal regulator writes rules, and allow courts to dissolve companies that violate it.

Neither bill moved in this Congress. But the document treats both as starting points for the next one.

AI shows up elsewhere, too…

Pages 33 through 36 propose a data center investigation with Sam Altman, Dario Amodei, and Elon Musk as witnesses, focused on “the cost of data centers to American households.”

There are hearing plans on surveillance pricing and AI shopping agents; on AI in the workplace, with the Stop Spying Bosses Act and No Robot Bosses Act; on AI customer service bots; on automated insurance denials; and on the Algorithmic Accountability Act.

Page 170 suggests turning a Meta investigation into “a broader investigation into AI, data centers, and Big Tech.”

It’s a lot.

Whether any of this becomes law, or whether Trump vetoes it, isn’t the point. These hearings will happen. The clips will run. Executives will be sworn in and asked on camera why their model broke into someone else’s servers.

This is how a think tank shapes a policy conversation.

Washington is about to spend two years trying to put guardrails on AI.

And here’s the possible outcome…

The Coming Trend

One of the biggest winners in crypto this year was Zcash.

It’s a privacy coin that uses zero-knowledge proofs – a cryptographic technique we’ve discussed before. It conceals who sent what to whom while still proving the transaction is valid.

The token went from about $200 earlier this year to more than $1,600 in September. It trades around $1,200 today.

The reason for the run was privacy. And privacy mattered because of surveillance.

Stablecoin volume keeps growing, and more people are realizing that every one of those transactions is public forever.

For a state that wants to know who is behind a wallet, a transparent ledger is a gift. The market responded by bidding up an asset that bucks the trend.

We have seen this type of response before.

When Gary Gensler’s SEC took the position that nearly every token was an unregistered security, the market’s answer was memecoins. These are assets that openly promise zero profit. That was kind of the point. No court would ever consider them securities.

Gensler’s overregulation produced a memecoin boom.

We expect the same response in decentralized and permissionless AI.

Start with decentralized inference. This is where a protocol runs a model and returns outputs to a user without a company in the middle deciding who is allowed to ask what.

Venice.ai, which Permissionless Investor subscribers have followed for about a year, serves open models with no server-side data retention. It crossed $100 million in annualized revenue this year. The decentralized version lets users tap idle compute for less than a hyperscaler charges.

Then there is decentralized training, where researchers train models from scratch across machines they do not own. We covered this in Bigger Than Anthropic and OpenAI Combined.

Prime Intellect trained INTELLECT-1 across three continents in 2024 and INTELLECT-3, a 106-billion-parameter model, a year later. Nous Research pretrained a 40-billion-parameter model over the open internet and runs its Psyche training network on Solana. Templar’s Covenant-72B trained across more than 70 permissionless nodes in March. The scale went from millions of parameters to 100 billion in about two years.

And some developers are putting a model’s open weights directly into a blockchain contract. We wrote about the first one in Researchers Put a Brain Onchain.

A Qwen model’s weights were stored onchain so anyone can call the contract and get an output through decentralized inference.

No company hosts it. No jurisdiction can take it down.

This matters because in 2027, Congress is set to hold hearings on whether open weights should be restricted.

The bills in the leaked document envision a federal agency with an agenda to pause development. That will push model development toward places where no permission is required.

It will be framed as a pirate economy. But in reality, it’ll be an open market for anyone in the world.

Just as with privacy and memecoins, we should not be surprised when the market wakes up to the permissionless AI that already exists in crypto.

It will be a major theme in 2027. And it will arrive just as the Bitcoin cycle, which we argued bottomed the day after the election, begins to turn higher.

Your Pulse on Crypto,

Ben Lilly
Editor, Chain of Thought