#330: Is It Possible That Your Body Will Heal By Creating And Manufacturing Its Own Medicine?, & More

The rapid cost-declines associated with gene editing and gene synthesis are turbo-charging scientific creativity, enabling the testing of hypotheses that, until now, have been cost-prohibitive. Recently, scientists at Nationwide Children’s Hospital published a proof-of-concept study demonstrating that gene therapies might deliver cancer immunotherapies more effectively than the current standard of care.

While gene therapies typically fix or replace faulty DNA, Nationwide Children’s Hospital’s DNA instructions encoded a protein-based immunotherapy for leukemia called Blincyto that the Federal Drug Administration (FDA) already had approved. Using novel DNA instruction-coding, the researchers enabled mouse cells to manufacture Blincyto continuously. Bundled into the therapy were DNA instructions for an ON/OFF switch that controlled the production of Blincyto.

Should this early research enter later stage clinical trials, the impact could be profound. Like Blincyto, many drugs require continuous infusions, holding patients hostage to IVs (intravenous therapies) in the hospital or at home. Nationwide’s research suggests that a single dose of gene therapy could induce the continuous production of medicine inside the body for up to 36 weeks.

Although this technique could face some of the delivery and targeting drawbacks associated with other gene therapies,[1] Nationwide’s breakthrough research could impact the administration of simpler, protein-based drugs like Keytruda or Opdivo.

[1] It is possible that the FDA will consider this a gene therapy even though the therapeutic agent is an immunotherapy.

Google researchers are creating home robots that can respond to vague language commands and solve physical problems. While prior robots required specific instructions like “Pick up the energy bar,” the new “helpers” respond to text prompts like “I’m tired. Bring me a snack that’ll give me some energy, please,” thanks to PaLM, Google’s language model. Along with chain-of-thought prompting and an affordance function that constrains outcomes, PaLM has improved robot algorithms enough to understand complex human language and, in contrast to “black box” deep learning models, create answers and solutions.

Conceptually, language-based, step-by-step reasoning could enable autonomous cars. Imagine, for example, a local police station prompting an autonomous taxi: “Pull over and let the passenger out of the vehicle.” A system incorporating chain-of-thought prompting––and step-by-step reasoning––could accelerate regulatory compliance and approvals.

In last week’s newsletter, ARK discussed how the US Treasury Department’s Office of Foreign Assets Control (OFAC) sanctions on Tornado Cash set a precedent that could be problematic for cryptocurrency privacy controls.

Since then, the sanctions have sparked fear about the sustainability of Ethereum’s censorship resistance at the base layer, particularly in the face of Ethereum’s transition to proof-of-stake validation. Proof-of-stake could expose ~65% of validators to US-based regulated companies that will have to comply with OFAC sanctions. As of Friday, the companies that dominate the Beacon Chain validators include Lido, Coinbase, Kraken, Staked.US, and Bitcoin Suisse.

Debates focused on preventing the censorship of Ethereum have settled on several scenarios and their tradeoffs. Widely supported is the user-activated soft fork that would remove Coinbase and other validators surgically from the network based on a supermajority of nodes, or users. In this scenario, entities censoring transactions could lose billions of dollars and face a contentious hard fork, splitting the network into an OFAC-compliant censorable chain and a non-OFAC-compliant “uncensorable” chain. In response, centralized entities like Circle or Coinbase might adopt the OFAC-compliant chain, impairing the non-compliant chain’s DeFi ecosystem.

Last week, CEO Brain Armstrong stated publicly that, if Coinbase had no way to challenge network-wide censorship, he believed it would likely shut down its Ethereum staking operations to preserve network integrity. As one of the largest ETH 2.0 staking pools in the network, Coinbase could become trapped between a rock and a hard place in the face of OFAC sanctions, potentially compromising crypto’s ethos or un-staking its ETH from the Ethereum network prematurely and losing billions of dollars in revenue and customer funds.