Lifeboat Foundation

Artificial intelligence (AI) has introduced a dynamic shift in various sectors, most notably by deploying autonomous agents capable of independent operation and decision-making. These agents, powered by large language models (LLMs), have significantly broadened the scope of tasks that can be automated, ranging from simple data processing to complex problem-solving scenarios. However, as the capabilities of these agents expand, so do the challenges associated with their deployment and integration.

Within this evolving landscape, a major hurdle has been the efficient management of LLM-based agents. The primary issues revolve around allocating computational resources, maintaining interaction context, and integrating agents with varying capabilities and functions. Traditional approaches often lead to bottlenecks and underutilization of resources, undermining these intelligent systems’ potential efficiency and effectiveness.

A research team from Rutgers University has developed the AIOS (Agent-Integrated Operating System), a pioneering LLM agent operating system designed to streamline the deployment and operation of LLM-based agents. This system is engineered to enhance resource allocation, enable the concurrent execution of multiple agents, and maintain a coherent context throughout agent interactions, optimizing agent operations’ overall performance and efficiency.

Generative models for chemical structures are often trained to create output in the common SMILES notation. Michael Skinnider shows that training models with the goal of avoiding the generation of incorrect SMILES strings is detrimental to learning other chemical properties and that allowing models to generate incorrect molecules, which can be easily removed post hoc, leads to better performing models.

Ever wonder where all the active supermassive black holes are in the universe? Now, with the largest quasar catalog yet, you can see the locations of 1.3 million quasars in 3D.

The catalog, Quaia, can be accessed here.

“This quasar catalog is a great example of how productive astronomical projects are,” says David Hogg, study co-author and computational astrophysicist at the Flatiron Institute, in a press release. “Gaia was designed to measure stars in our galaxy, but it also found millions of quasars at the same time, which give us a map of the entire universe.” By mapping and seeing where quasars are across the universe, astrophysicists can learn more about how the universe evolved, insights into how supermassive black holes grow, and even how dark matter clumps together around galaxies. Researchers published the study this week in The Astrophysical Journal.

Race Speeds Up. Wallets open up. Agi 2025–2029.

Amazon is reportedly planning to spend a whopping $150 billion within the next 15 years on building data centers — a move that will position the tech giant to be able to handle an expected explosion with artificial intelligence applications and other digital services.

The spending spree, earlier reported on by Bloomberg, will also allow Amazon to maintain its top spot in the cloud services market, where it holds roughly twice the share of No. 2 player Microsoft.

Continue reading “Amazon reportedly to spend $150B to build data centers needed for AI boom, ‘get closer to customers’” »

Is anything ever normal in the AI industry?

A doozy of a scoop by the newsletter Nongaap Investing and extensively followed up by Business Insider certainly makes us wonder. The gist is that for a period of time in 2023, the person in charge of OpenAI’s $175 million startup fund appears to have been completely fake — and OpenAI says the documents filed with the California Secretary of State to put the fake person in charge were “completely fabricated.”

Head spinning yet? Us too. OpenAI is almost certainly the hottest startup on the planet right now, and it sounds like someone pulled an extrardinary fast one on it, with unclear goals. And lest you think this is some unimportant position, the person now running the fund is none other than OpenAI’s mercurial CEO, Sam Altman.

Medical implants such as pacemakers and gastric stimulators have improved our lives, but the batteries in these devices eventually run out and require surgery to replace them.

It raises a futuristic question: what if there was a way to avoid cutting a patient’s body open to replace a battery?

A team of Chinese scientists have come up with a possible method to pull that off by developing an implantable battery that uses oxygen already inside the human body to continuously power itself up.

#jwst Subscribe to Science Time: https://www.youtube.com/sciencetime24Div…

How fast can a neutron star drive powerful jets into space? The answer, it turns out, is about one-third the speed of light, as our team has just revealed in a new study published in Nature.

In an interview, University of Michigan civil engineer Sherif El-Tawil explained how often ships collide with bridges, what can be done to protect bridges from collisions, and how a similar disaster in Florida in 1980 – just three years after the Key bridge opened – changed the way bridges are built.

This is not the first time a ship has taken out a bridge. What’s the history of ship-bridge collisions?

This is an extremely rare event. To my knowledge, there are about 40 or so recorded events in the past 65 years that involved similar type of damage to a bridge caused by a ship. So they seem to occur on average about once every one and a half to two years around the world. When you consider that there are millions of bridges around the world – and most of them cross waterways – you can imagine how rare this is.