OpenAI has announced ‘The Stargate Project,’ a new company set to invest $500 billion into AI infrastructure over the next four years
The data centers will be exclusively used by OpenAI as it expands its generative AI compute portfolio. Of the total investment, $100bn will be deployed ‘immediately.’
SoftBank, OpenAI, Oracle, and Abu Dhabi’s MGX are the equity investors in Stargate, with SoftBank having financial responsibility and OpenAI having operational responsibility. SoftBank’s Masayoshi Son will serve as chairman.
The buildout is currently underway, starting in Texas – likely Oracle’s project in Abilene, Texas, which is itself leased from Crusoe. — Read More
Recent Updates Page 84
BrowserAI
Deepseek: The Quiet Giant Leading China’s AI Race
Deepseek is a Chinese AI startup whose latest R1 model beat OpenAI’s o1 on multiple reasoning benchmarks. Despite its low profile, Deepseek is the Chinese AI lab to watch.
… Deepseek’s strategy is grounded in their ambition to build AGI. Unlike previous spins on the theme, Deepseek’s mission statement does not mention safety, competition, or stakes for humanity, but only “unraveling the mystery of AGI with curiosity”. Accordingly, the lab has been laser-focused on research into potentially game-changing architectural and algorithmic innovations.
Deepseek has delivered a series of impressive technical breakthroughs. Before R1-Lite-Preview, there had been a longer track record of wins: architectural improvements like multi-head latent attention (MLA) and sparse mixture-of-experts (DeepseekMoE) had reduced inference costs so much as to trigger a price war among Chinese developers. Meanwhile, Deepseek’s coding model trained on these architectures outperformed open weights rivals like July’s GPT4-Turbo.
As a first step to understanding what’s in the water at Deepseek, we’ve translated a rare, in-depth interview with CEO Liang Wenfeng, originally published this past July on a 36Kr sub-brand. — Read More
AI Will Write Complex Laws
Artificial intelligence (AI) is writing law today. This has required no changes in legislative procedure or the rules of legislative bodies—all it takes is one legislator, or legislative assistant, to use generative AI in the process of drafting a bill.
In fact, the use of AI by legislators is only likely to become more prevalent. There are currently projects in the US House, US Senate, and legislatures around the world to trial the use of AI in various ways: searching databases, drafting text, summarizing meetings, performing policy research and analysis, and more. A Brazilian municipality passed the first known AI-written law in 2023.
That’s not surprising; AI is being used more everywhere. What is coming into focus is how policymakers will use AI and, critically, how this use will change the balance of power between the legislative and executive branches of government. Soon, US legislators may turn to AI to help them keep pace with the increasing complexity of their lawmaking—and this will suppress the power and discretion of the executive branch to make policy. — Read More
Deep-learning enabled generalized inverse design of multi-port radio-frequency and sub-terahertz passives and integrated circuits
Millimeter-wave and terahertz integrated circuits and chips are expected to serve as the backbone for future wireless networks and high resolution sensing. However, design of these integrated circuits and chips can be quite complex, requiring years of human expertise, careful tailoring of hand crafted circuit topologies and co-design with parameterized and pre-selected templates of electromagnetic structures. These structures (radiative and non-radiative, single-port and multi-ports) are subsequently optimized through ad-hoc methods and parameter sweeps. Such bottom-up approaches with pre-selected regular topologies also fundamentally limit the design space. Here, we demonstrate a universal inverse design approach for arbitrary-shaped complex multi-port electromagnetic structures with designer radiative and scattering properties, co-designed with active circuits. To allow such universalization, we employ deep learning based models, and demonstrate synthesis with several examples of complex mm-Wave passive structures and end-to-end integrated mm-Wave broadband circuits. The presented inverse design methodology, that produces the designs in minutes, can be transformative in opening up a new, previously inaccessible design space. — Read More
Google DeepMind CEO: AI-Designed Drugs Coming to Clinical Trials in 2025
Nobel laureate and Google DeepMind CEO Demis Hassabis said Tuesday (Jan. 21) that he expects to see pharmaceutical drugs designed by artificial intelligence (AI) to be in clinical trials by the end of the year.
During a fireside chat at the World Economic Forum in Davos, Switzerland, Hassabis said these drugs are being developed at Isomorphic Labs, a for-profit venture created by Google parent firm Alphabet in 2021 that was tasked to reinvent the entire drug discovery process based on first principles and led by AI.
“That’s the plan,” Hassabis said. — Read More
AI Mistakes Are Very Different from Human Mistakes
Humans make mistakes all the time. All of us do, every day, in tasks both new and routine. Some of our mistakes are minor and some are catastrophic. Mistakes can break trust with our friends, lose the confidence of our bosses, and sometimes be the difference between life and death.
Over the millennia, we have created security systems to deal with the sorts of mistakes humans commonly make. … Humanity is now rapidly integrating a wholly different kind of mistake-maker into society: AI. Technologies like large language models (LLMs) can perform many cognitive tasks traditionally fulfilled by humans, but they make plenty of mistakes. It seems ridiculous when chatbots tell you to eat rocks or add glue to pizza. But it’s not the frequency or severity of AI systems’ mistakes that differentiates them from human mistakes. It’s their weirdness. AI systems do not make mistakes in the same ways that humans do.
Much of the friction—and risk—associated with our use of AI arise from that difference. We need to invent new security systems that adapt to these differences and prevent harm from AI mistakes. — Read More
Evolving Deeper LLM Thinking
We explore an evolutionary search strategy for scaling inference time compute in Large Language Models. The proposed approach, Mind Evolution, uses a language model to generate, recombine and refine candidate responses. The proposed approach avoids the need to formalize the underlying inference problem whenever a solution evaluator is available. Controlling for inference cost, we find that Mind Evolution significantly outperforms other inference strategies such as Best-of-N and Sequential Revision in natural language planning tasks. In the TravelPlanner and Natural Plan benchmarks, Mind Evolution solves more than 98% of the problem instances using Gemini 1.5 Pro without the use of a formal solver. — Read More
#performanceChina to host world’s first human-robot marathon as robotics drives national goals
For the first time, dozens of humanoid robots are expected to join a half-marathon to be held in the capital’s Daxing district in April, according to local authorities.
This comes as China ramps up efforts to develop artificial intelligence and robotics, to gain an edge in the tech rivalry with the US as well as combat the challenges of an ageing society and a falling birth rate.
Some 12,000 humans will take part in the coming race – and running alongside them on the 21km (13-mile) route will be robots from more than 20 companies, according to the administrative body of Beijing Economic-Technological Development Area, or E-Town.
Prizes will be offered for the top three runners. — Read More
Rick's Cafe AI 