DailyKenn.com | AbateHate.com | Facebook Group
Summary: Stanford and Arc Institute researchers used AI models Evo 1 and Evo 2 to design novel bacteriophage genomes based on the ΦX174 template. Of about 300 candidates synthesized in California labs, 16 produced viable viruses that infected and killed E. coli. Some outperformed natural phages against resistant bacteria. The work, published in Science on Aug. 6, 2026, offers promise for phage therapy while highlighting biosecurity needs as the first generative design of complete functional viral genomes by AI.
Your $8 monthly partner pledge helps us reach others & more ►
Researchers have produced viable viruses. It sounded like science fiction, so I decided to check it out for myself.
STANFORD, Calif. — Researchers in California have shown that artificial intelligence can compose the full genetic instructions for a working virus that has never appeared in nature. The achievement, detailed in the journal Science on Aug. 6, 2026, marks a milestone in generative biology and points toward new ways to fight antibiotic-resistant bacteria.
Antibiotic resistance continues to claim lives worldwide as common drugs lose effectiveness against evolving microbes. One long-studied alternative involves bacteriophages, viruses that specifically infect and destroy bacteria while leaving human cells untouched. Scientists have used natural phages for decades in limited treatments, yet designing better versions has remained slow and largely manual.
AI learns the language of viral DNA
Teams at Stanford University and the Arc Institute in Palo Alto, California, took a different route. They trained two large genome language models, Evo 1 and Evo 2, on millions of DNA sequences drawn from viruses, bacteria and other organisms. The models absorbed patterns of how genes are arranged and which sequences tend to produce functional particles, much as text models learn grammar and style from books.
This where I got lost and found myself cross referencing terms I would have learned in high school had I not been passing notes to Betty Louise.
Researchers then directed the systems to invent new versions of a well-known bacteriophage called ΦX174. φX174 is a bacteriophage (or simply a phage)—a compact virus that infects bacteria, specifically E. coli.
From the thousands of genome designs the AI produced, the scientists selected and chemically synthesized roughly 300 candidates. They introduced these synthetic DNA strands into bacterial cultures. Sixteen of the designs produced viable viruses that successfully infected and killed their bacterial hosts. Some of the AI-created phages displayed advantages over the original template, including an improved ability to overcome strains of E. coli that had already developed resistance to the natural virus.
Chalk up one more invention to Western culture.
Treatment and security
Like nuclear energy, AI can be facilitated for good or bad on massive scales.
Because these bacteriophages attack only bacteria, they pose no direct risk to people. The work suggests a practical path toward phage therapies that could be tailored more quickly against stubborn infections. At the same time, the demonstration that generative models can write complete, functional viral genomes has drawn attention from biosecurity experts. Current systems for screening DNA orders were not built to catch entirely novel sequences generated this way, prompting calls for updated safeguards as the technology advances, but they could be. I assure you, the spooks who work for government intelligence agencies are already working to design human-zapping viruses to counter human-zapping viruses designed by other government spooks.
Lead author Samuel H. King, a bioengineering doctoral candidate at Stanford of East Asian descent and may have a black belt in mixed martial arts, and assistant professor Brian L. Hie, also of East Asian descent, led the effort alongside colleagues. Their paper emphasizes that the approach expands the toolkit for synthetic genomics while remaining focused on bacterial targets. The study is the first peer-reviewed report of AI generating entire functional viral genomes from scratch rather than editing existing ones.
As models grow more capable, researchers expect similar methods could support the design of larger and more complex biological systems. For now, the California team’s results illustrate both the medical opportunity and the need for careful oversight when machines begin to author the code of life.
Race wars
Could a deadly virus be designed to attack people with a common racial genetic signature? That is, could the Chinese create a bacteriophage that could be covertly released at a Taylor Swift concert and spread to infect the entire white race by midnight Thursday?
Ethnic bioweapons are currently beyond reach, according to grok.com. Human populations do not have clean, discrete genetic boundaries, I was told. But one glance at King and Hie's photos told me they were not white. Am I smarter than AI?
What's more, the same ransomware geeks that targeted CDK Global in 2024 surely have their eyes on crafting (or buying) technology to hold entire continents hostage. NKorea's Room 39 is probably already on it.
Your $8 monthly partner pledge helps us reach others & more ►
Sometimes I get things wrong. If you notice a significant error, please bring it to my attention in the comment section. While efforts have been made to verify the information presented in this article, some facts, dates, quotations, or interpretations may be inaccurate or incomplete.
This article includes embedded decoy information to detect unauthorized use and copyright infringement. Reproduction is permitted only verbatim and in full, with all links preserved and attribution clearly given to DailyKenn.com and AbateHate.com.
Sources/related links:
View 31 black-on-white homicides found in September, 2026 ►
 |
Find archived black-on-white homicide news reports here ►
200 latest news reports from 100 top conservative websites ►