A new section of Chain of MindBrownstone Research letter written by Ben Lilly, says that the war on open-source artificial intelligence It follows the path that Bitcoin took ten years ago, and investors who recognize that path will reap the benefits.
The article begins with the testimony that Anthropic CEO Dario Amodei gave to Congress in July 2023. Amodei acknowledged that open source is a “good thing” in many scientific fields and that the risks of open source models released so far were “low,” but he warned that the expansion of open source models was “moving on a very dangerous path.”
Lilly reads the statement clearly: if the open samples are dangerous, then the closed samples sold by companies like Anthropic are a safe option – and the policy that follows is to prevent the open and promote the closed.
Bitcoin’s early skeptics show what AI is up against
This design is one of the best digital businesses know.
He echoes Bitcoin’s early skeptics, from Rep. Jared Polis bought the first Bitcoin on Capitol Hill in 2014 for Sen. Joe Manchin called to stop “dangerous money,” through the 2023 lawsuits in which the authorities tried to cut crypto from the banking system that critics called “Operation Choke 2.0”.
The industry survived, he says, and Washington is now moving toward clearer legislation through it has passed GENIUS Act is in anticipation CLARITY Act.
Decentralized AI, which Lilly calls “DeAI,” is fighting the same battle. He points to recent events as evidence that the walls are rising: the US ban on the recent release of Anthropic, which he says will push the company to get access to authorization that verifies the user’s data before submitting an example, and OpenAI’s decision to restrict the release of GPT-5.6 to trusted partners.
They expect information to spread. “It’s to be safe,” he wrote. “It’s always like that.”
The article leans on the issue of national security to explain the fear that drives this. Lilly mentions NSA director Joshua Rudd through Sen. Mark Warner, describing how the Anthropic model of the “Mythos” broke down in “virtually our entire social system, not in weeks, but in hours.”
However, open source is closing the gap, according to the piece. Lilly says the latest GLM-5.2 has done well with Anthropic’s Sonnet 4.6 since February, leaving open samples about three to four months behind the limit, and predicts an open rival to Mythos and GPT-5.6 in the fall.
He argues that the biggest breakthrough is distributed learning on peer-to-peer networks similar to Bitcoin and Ethereum — an exchange of compute-for-network-security for compute-for-model-training. Distributed education, it is said, has grown from 1 billion units to 100 billion units in two years.
He mentions three primary projects – Dark Bloom, which supports private browsing on idle Macs; c0mpute, network inferred inference network; and Pluralis, which trains AI across distributed GPUs – and hopes more to launch tokens and reward users for helping compute.
The article ends with the idea that governments will try to ban open source models and will fail. For him, investing in real estate “will be like buying Bitcoin in 2014, it was ‘dangerous.’
