It stands out beyond the spectrum of artificial intelligence use.
I agree that the great and exciting AI opportunities on the horizon are the acceleration and transformation of scientific discovery and development. Supplied to a huge crowd of scientific data, AI is committed to creating new medicines to combat disease, the world’s population, and new ingredients, in order to unleash green energy.
Technology companies such as Microsoft and Google have created AI tools for science and are working with partners in areas such as Drug Discovery. And last year, the Nobel Prize in Chemistry went to scientists using AI to predict and create proteins.
This month, Lila Sciences has its own ambition to revolutionize science through AI based in Cambridge, Massachusetts, working secretly for two years “to build scientific tensions to solve humanity’s greatest challenges.”
Lila relies on an experienced team of scientists and initial $200 million funding to develop an AI program trained on published experimental data and scientific processes and inference. With the startup, AI software has several scientists who will help run experiments in automated physical labs.
Already in a project demonstrating the technology, Lila’s AI has developed new materials for generating new antibodies to combat disease and capturing carbon from the atmosphere. Lila transformed these experiments into physical results in the lab within a few months. This is a process that can likely take years to traditional research.
Experiments like Lira have convinced many scientists that AI will soon make the hypothetical experimental test cycle faster than ever. In some cases, AI can even surpass the human imagination with the progression of inventions, turbochargers.
“AI drives the next revolution of this most precious thing that humans have ever stumbled,” said Geoffrey von Maltzahn, CEO of Lila. in biomedical engineering and medical physics from Massachusetts Institute of Technology.
The push to reinvent the scientific discovery process is based on the power of generator AI and exploded into public awareness more than two years ago with Openai’s introduction of ChatGPT. New technologies are trained on data across the Internet, allowing you to answer questions, write reports, and create emails in human-like flow ency.
The new AI has launched a commercial arms race and seemingly endless spending by high-tech companies such as Openai, Microsoft, and Google.
(The New York Times sued Openai and Microsoft to form a partnership and condemned copyright infringement over news content related to AI systems. Openai and Microsoft denied these claims.)
Lila employs a science-focused approach that trains generative AI, supplies IT research papers, and documents experiments and data from the growing Life Science and Materials Science Lab. The Lila team believes that chatbots will give technology both science and a wide range of capabilities, reflecting the way chatbots write poetry and computer code.
Still, companies working to decipher the lira and “scientific tensions” will face major challenges, scientists say. AI is already revolutionizing certain areas, including drug discovery, but it is unclear whether this technology is merely a powerful tool, or whether it matches or outweighs all human capabilities.
Because lira is run secretly, outside scientists cannot evaluate their work, and early advances in science do not guarantee success.
“We’ve seen a lot of different things,” said David Baker, director of the University of Washington’s biochemist and protein design research. “It seems like it’s beyond what I know very well in scientific discoveries.”
Dr. Baker, who shared the Nobel Prize in Chemistry last year, said he viewed AI as a more tool.
Lila was conceived within the flagship pioneer, an investor and prolific creator of biotech companies, including Covid-19 vaccine maker Moderna. Flagship is conducting scientific research that can prove commercially valuable, focusing on where breakthroughs could be within a few years.
“So we don’t only care about this idea, we also care about the timeliness of the idea,” Dr. Afiyan said.
Lila focused on new materials with the merger of two early AI enterprise projects in the flagship, and the other focused on biology. The two groups were trying to solve similar problems and recruit the same people, so they combined the powers, said Molly Gibson, a computational biologist and co-founder of Lira.
The LILA team has completed five projects to demonstrate the capabilities of AI. This is a powerful version that is one of the sophisticated assistants known as agents. In either case, scientists who were not expert in the subject were usually entered in requests that they wanted to achieve in the AI program. After improving the request, the scientists worked as partners with AI to run the experiment and test the results.
One of these projects discovered a new catalyst for green hydrogen production. This involves using electricity to divide water into hydrogen and oxygen. AI was instructed that unlike current commercial standard, iridium, the catalyst must be rich or easy. With the help of AI, two scientists found a new catalyst in four months. This can take more generally several years.
Its success convinced John Gregoire, a well-known researcher of new materials for clean energy, to leave California Institute of Technology last year, and to leave Lira as head of physical science research.
George Church is a Harvard geneticist known for his pioneering research into genomic sequencing and DNA synthesis, which co-founded dozens of companies, and recently joined as Lira’s Chief Scientist.
“I think science is a really good topic for AI,” Dr. Church said. Science focuses on specific areas of knowledge, and he adds that truth and accuracy can be tested and measured. This reduces the tendency for false and false answers known as hallucinations, which can also be created by chatbots.
Early projects still have a long way from market-ready products. Lila works with partners to commercialize ideas born from the lab.
Lila expands its lab space in the sixth floor flagship building in Cambridge alongside the Charles River. Over the next two years, Lila says she plans to move to another building and add tens of thousands of square feet of lab space and open offices to San Francisco and London.
On a recent day, trays carrying 96 DNA samples rode on magnetic tracks, shifting directions quickly for delivery to different lab stations, in part to what AI proposed. This technique appeared to be improvisational as it performed experimental steps in the pursuit of novel proteins, gene editors, or metabolic pathways.
In another part of the lab, scientists monitored high-tech machines used to create, measure and analyze custom nanoparticles for new materials.
Activities on the lab floor were guided by a collaboration of white-coated scientists, automated equipment and invisible software. All measurements, all experiments, all progressive successes and failures were captured digitally and fed to Lila’s AI, allowing you to learn continuously, become smarter and do more on its own.
“Our goal is to provide AI access to implement scientific methods. To come up with new ideas, then actually enter the lab and test those ideas,” Dr. Gibson said.
