Citrini and Anthropic test Claude’s design of scorpion antivenom proteins; 17 candidates can bind to the target toxin

AT A GLANCE

Citrini used Claude to design scorpion antivenom proteins and, based on experimental validation bottlenecks, is focusing on outsourced research and development, testing equipment, and automation tools.

Article

「AI Publisher of the year-end report” Citrini had Claude research scorpion venom to seek new investment opportunities

Beating AI breaking news, this year February, Citrini Research published “2028 Global Intelligence Crisis,” envisioning AI massively replacing white-collar workers and triggering an economic recession, which once roiled Wall Street. Recently, this investment research firm partnered with Anthropic to have Claude attempt to design scorpion antivenom proteins.

Citrini began seeking June from AI infrastructure this year, focusing on the long-depressed life sciences sector. However, experts are sharply divided over the prospects of AI pharmaceuticals, and 5 researchers with no drug development experience decided to verify it themselves. They chose scorpion venom because its toxin structure is relatively clear, making experiments easier to conduct. Moreover, traditional antivenom serum requires injecting horses with venom and then extracting antibodies, making the production process cumbersome and costly.

The team obtained Anthropic’s not-yet-public research model. With researchers’ assistance, Claude spent 64 hours generating approximately 320000 protein designs and selected 180 candidates. It then used Claude to add 31 more, which were sent to two laboratories for testing. Ultimately, 17 proteins were able to bind to the target scorpion venom, but none could bind multiple toxins simultaneously, and it has not yet been verified whether they can actually neutralize the venom.

What truly interested Citrini was the follow-up experiment.AI completed the designs in just a few days, while laboratory validation took nearly 2 months. Computing costs were approximately 50000 USD, borne by Anthropic; the laboratory bill, meanwhile, was close to 150000 USD.

Citrini thereby identified a new investment opportunity. As AI accelerates drug design, more candidate molecules may need to be sent to laboratories for synthesis, testing, and validation. These tasks cannot be completed by computers alone; they also require expensive equipment, reagents, and specialized personnel.

Therefore, Citrini is not only focused on companies using AI to develop drugs, but also favors Charles River, Revvity, and Tecan, which provide pharmaceutical companies with outsourced laboratory services, testing equipment, and automation tools.

Original link https://m.theblockbeats.info/flash/373240

Key points

This year February, Citrini Research published “2028 Global Intelligence Crisis,” envisioning AI massively replacing white-collar workers and triggering an economic recession.
Citrini partnered with Anthropic to have Claude attempt to design scorpion antivenom proteins.
Claude spent 64 hours generating approximately 320000 protein designs and selecting 180 candidates, then added 31 more and sent them to two laboratories for testing.
Testing showed that 17 proteins could bind to the target scorpion venom, but none could bind multiple toxins simultaneously, and it has not yet been verified whether they can actually neutralize the venom.
AI The designs took a few days, while laboratory validation took nearly 2 months; approximately 50000 USD in computing costs was borne by Anthropic, and the laboratory bill was close to 150000 USD。
Citrini favors Charles River, Revvity, and Tecan for providing pharmaceutical companies with outsourced laboratory services, testing equipment, and automation tools.

AI-assisted interpretation

The following is analysis, separate from reported facts. Verify important claims independently.

This attempt shows that AI a large number of protein designs can be proposed quickly, but candidate designs still need to be synthesized, tested, and validated in the laboratory. The existing results only demonstrate that some proteins can bind to the target scorpion venom; this does not mean that an effective antivenom has already been developed.

Why it matters to readers

Based on this, Citrini believes that if AI causes the number of drug candidates to increase, demand for outsourced laboratory services, testing equipment, reagents, and automation tools may also increase. However, the original article provides no information on changes in orders or performance at the relevant companies.

Beginners should distinguish between “AI generating candidate designs,” “protein-toxin binding,” and “actually neutralizing venom.” This message currently mainly reflects the experimental process and Citrini’s investment view, rather than indicating that the drug has already succeeded.

Risks and unknowns

Whether the candidate proteins can actually neutralize venom has not yet been validated.
No candidate protein has been shown to bind multiple toxins simultaneously.
The original text does not provide raw experimental data, details of the research model, or formal research materials.
AI It remains uncertain whether drug design will lead to actual orders or revenue growth for experimental services and equipment companies.

Citrini and Anthropic test Claude’s design of scorpion antivenom proteins; 17 candidates can bind to the target toxin

Article

AI-assisted interpretation

Why it matters to readers

Risks and unknowns

Related Developments

Related concepts

Keep following

Security Center