Future houseAn Eric Shmidt-backed non-profit with the goal of creating an “AI scientist” in the next decade has introduced its first major product: a platform and API designed to support scientific work.
Many, many startups are racing to develop AI research equipment for scientific domains, some have a lot of VC funding behind them. Tech giants also seem to be a bullshit in AI for science. Earlier this year, Google has unveiled ”AI co-scientist“The agency says an AI has said that scientists can help create assumptions and experimental research plans.
AI companies have CEO OpenAI and anthropologists Seer AI equipment can greatly accelerate scientific discoveries, especially in treatment. However, many researchers do not think that AI is particularly effective in guiding the scientific process, because of its incredible part.
The Future House released four AI equipment on Thursday: Crow, Falcon, Owl and Phoenix. The crow can search scientific literature and answer the questions about it; Falcon can handle deep literature search with scientific database; Owl looks for previous work in a particular subject region; And the Phoenix Chemistry Test uses tools to help in the audit plan.
“Not the same as someone else [AIs]The Future House has access to a huge corpus of high-quality open-access papers and specialized scientific tools, ” Wrote Future House on a blog post. “They are [also] There is a transparent argument and use a multi-paid process to consider each source more depth […] By this discipline [AI]Together, scale, scientists can accelerate the pace of scientific discoveries. “
Obviously, Futurehouse has not yet discovered a scientific progress or discover a novel with its AI equipment.
An “AI scientist” is a part of the challenge of developing the development of a countless number of misleading factors. AI may be effective in areas where a huge list of possibilities are needed to narrow. However, it is less clear whether the AI is capable of solving the problem of boxes, whether it leads the Bonafide breakthroughs.
TechCrunch event
Berkeley, CA
|
June 5
Book now
The results of the AI systems designed for science are still degraded in most cases. Google said in 2023 that there were about 40 new materials Synthesized One of these AIS is known as genome. Yet An outer analysis None of the materials were found, in fact the net is new.
AI’s technical defects and risks tendency to hallucinet, scientists warn it about supporting it for serious work. Even well-known studies can end by abusing AI, which fights to perform high-dependency work.
In fact, the Future House admits that its AI equipment – especially Phoenix – can make mistakes.
“We are releasing [this] Now in the consciousness of a quick repetition, “The company wrote in his blog post.” Provide feedback as soon as you use it. “