#InfocusAI is rushing to you with another batch of news about artificial intelligence. In this issue, we will talk about batteries for microrobots, testing LLMs’ abilities in inductive and deductive reasoning, a new forecast from IDC regarding spending on AI, Procreate’s rejection of generative technologies and Hugging Face’s next step towards democratization of robotics.
AI-focused digest – News from the AI world
Issue 48, August 9-22, 2024
MIT has developed a zinc-air battery for tiny autonomous robots
Engineers from MIT have developed a tiny (smaller than a grain of sand) zinc-air battery for microrobots. As reported on MIT News, the new invention is only 0.1 millimeters long and 0.002 millimeters thick and is capable of creating a current with a potential of up to 1 volt, capturing oxygen from the air and oxidizing zinc with it. This is enough to power a small circuit, sensor, or actuator. The built-in zinc-air battery will give microrobots much greater autonomy compared to solar-powered systems that require an external light source to power them. Due to this, the scope of application of tiny robots will expand significantly. For example, they can be used to deliver drugs in the human body. See more about the invention on the pages of Science Robotics.
LLMs turned out to be strong in inductive reasoning and not particularly good at deduction
Large language models have impressive inductive reasoning abilities but are not too strong in deduction. It is especially difficult for them to cope with so called counterfactual tasks that deviate from the conventional cases seen during pre-training. This conclusion was made by scientists from the University of California, Los Angeles and Amazon as a result of a comprehensive study of LLMs’ abilities, in particular those of GPT-3.5 and GPT-4, in the application of certain types of reasoning. For their experiments, to separate induction from deduction and to study these types of reasoning in pure form in LLMs, they used a framework of their own development called SolverLearner. For details, go to this preprint.
Global AI spending to double by 2028
IDC predicts that global spending on artificial intelligence (including applications, infrastructure, and related IT and business services) will more than double by 2028 and reach $632 billion. The compound annual growth rate (CAGR) will be 29%. Despite the increased focus on generative AI, spending on this technology in the next five years will be less than the combined total of all other artificial intelligence technologies such as machine learning, deep learning, automatic speech recognition and natural language processing. However, thanks to significant investments in the development of generative technologies, the spending on solutions with generative AI will grow with a CAGR of 59.2% (this is faster than the market growth) and reach $202 billion, that is, almost a third of overall spending on artificial intelligence. See more figures in this IDC press release.
Hugging Face continues the democratization of robotics
Hugging Face has released an open access guide to the building and training of low-cost AI-powered robots. This is the next step of the open-source community upon the LeRobot initiative for the democratization of robotics, which is now dominated by large corporations and institutions with significant resources. The guide covers almost all stages of robot creation, i.e. from searching for parts to deploying AI models, so that anyone, regardless of their qualifications, can follow text and video instructions to create at least a robotic arm. VentureBeat tells you more about the project philosophy, and the guide itself is on GitHub.
Procreate has abandoned generative AI
The developers of the Procreate popular graphics editor for iPad have officially announced that the further development of their product will not be related to generative artificial intelligence. “AI is not our future”, this slogan they posted in large letters on their official website, explaining that they perfectly understand the risks of being left behind by following such a policy. “We’re here for the humans. We’re not chasing a technology that is a moral threat to our greatest jewel: human creativity”, the statement says. Rather boldly, we dare say. And what do you think about the rejection of generative AI?
