AI pioneer Francois Chollet leaves Google

Francois Chollet, a leading figure in the AI world, is leaving Google after close to a decade. In a post on X, Chollet said that he’s starting a new company with “a friend” — but declined to reveal additional details. “I’m very grateful for my decade at Google,” he said. “In that timespan, deep learning […]

© 2024 TechCrunch. All rights reserved. For personal use only.

Summary

Francois Chollet, a prominent AI researcher, has announced his departure from Google after nearly ten years. In a post on X, he mentioned that he will be starting a new company with a friend but did not provide further details. Chollet expressed gratitude for his time at Google, highlighting his contributions to deep learning during his tenure.

This article was summarized using ChatGPT

Prominent crypto critic says someone offered bribes to take down a blog post

Crypto critic Molly White received emails from a purported lawyer and someone working for a “reputation management company” offering bribes in exchange for the takedown of a blog post about the arrest of an alleged crypto fraudster.

© 2024 TechCrunch. All rights reserved. For personal use only.

Summary

Crypto critic Molly White reported receiving emails from a supposed lawyer and a representative from a “reputation management company” who offered bribes to have her blog post about the arrest of an alleged crypto fraudster removed.

This article was summarized using ChatGPT

OpenAI hires its first chief economist

OpenAI has hired its first chief economist: Aaron Chatterji, formerly the chief economist at the Commerce Department under President Joe Biden and a senior economist in President Barack Obama’s Council of Economic Advisers. Chatterji, who’s also a professor of business and public policy at Duke, will study AI’s economic impacts at OpenAI, leading research into how […]

© 2024 TechCrunch. All rights reserved. For personal use only.

Summary

OpenAI has appointed Aaron Chatterji as its first chief economist. Chatterji previously served as the chief economist at the Commerce Department under President Biden and was a senior economist in Obama’s Council of Economic Advisers. He is also a professor at Duke University and will focus on researching the economic impacts of AI at OpenAI.

This article was summarized using ChatGPT

Plugable’s new 10-port USB-C hub is exactly the overkill my nightstand needs

Plugable has announced a new multi-port USB-C hub that can power or charge up to 10 connected devices using intelligent power distribution and an easy-to-understand priority-based charging system.

The company is bringing its new PS-10CC hub to consumers through a Kickstarter crowdfunding campaign. The earliest backers can get one for $59, while the hub’s full MSRP will be closer to $99, with delivery expected as early as January 2025. There’s always a level of risk when backing a crowdfunded product, but Plugable has successfully delivered three devices through previous Kickstarters and says it already has functional prototypes of its new PS-10CC.

Continue reading…

Summary

Plugable has introduced the PS-10CC, a new 10-port USB-C hub that allows users to charge or power multiple devices using an intelligent power distribution system. The hub is being launched via a Kickstarter campaign, with early backers able to purchase it for $59, while the retail price will be around $99, with expected delivery by January 2025. Although crowdfunding carries some risks, Plugable has a solid track record of successfully delivering products and already has functional prototypes of the hub. Users will need to provide their own power adapter for the hub.

This article was summarized using ChatGPT

Researchers link Polyfill supply chain attack to huge network of copycat gambling sites

A supply chain hack targeting 100,000 websites was launched to redirect internet users to a massive online gambling network.

© 2024 TechCrunch. All rights reserved. For personal use only.

Summary

A supply chain attack affected 100,000 websites, redirecting users to a large network of copycat gambling sites. This incident highlights the vulnerabilities within the Polyfill supply chain.

This article was summarized using ChatGPT

This $600 Buzz Lightyear toy is the most realistic yet

Do you have $600 burning an asteroid-sized hole in your pocket? If so then the new Buzz Lightyear robot might be for you. A collaboration between Pixar and smart toy company Robosen, the new Buzz is fully loaded with over 3,000 tiny parts, 75 microchips, and 23 servo motors. It also features what Robosen calls […]

© 2024 TechCrunch. All rights reserved. For personal use only.

Summary

The article discusses a new Buzz Lightyear robot toy, priced at $600, created through a collaboration between Pixar and Robosen. This advanced toy includes over 3,000 parts, 75 microchips, and 23 servo motors, making it one of the most realistic Buzz Lightyear figures available.

This article was summarized using ChatGPT

Microsoft Open-Sources bitnet.cpp: A Super-Efficient 1-bit LLM Inference Framework that Runs Directly on CPUs

The rapid growth of large language models (LLMs) has brought impressive capabilities, but it has also highlighted significant challenges related to resource consumption and scalability. LLMs often require extensive GPU infrastructure and enormous amounts of power, making them costly to deploy and maintain. This has particularly limited their accessibility for smaller enterprises or individual users […]

The post Microsoft Open-Sources bitnet.cpp: A Super-Efficient 1-bit LLM Inference Framework that Runs Directly on CPUs appeared first on MarkTechPost.

Summary

Microsoft has open-sourced a new framework called bitnet.cpp, designed for efficient 1-bit large language model (LLM) inference that operates directly on CPUs. This development addresses the growing challenges associated with resource consumption and scalability of LLMs, which typically require expensive GPU infrastructure and significant power, making them less accessible for smaller enterprises and individual users. The bitnet.cpp framework aims to reduce these barriers, promoting broader usage of LLMs.

This article was summarized using ChatGPT

FusionANNS: A Next-Gen ANNS Solution that Combines CPU/GPU Cooperative Processing for Enhanced Performance, Scalability, and Cost Efficiency

Approximate nearest neighbor search (ANNS) is a critical technology that powers various AI-driven applications such as data mining, search engines, and recommendation systems. The primary objective of ANNS is to identify the closest vectors to a given query in high-dimensional spaces. This process is essential in contexts where finding similar items quickly is crucial, such […]

The post FusionANNS: A Next-Gen ANNS Solution that Combines CPU/GPU Cooperative Processing for Enhanced Performance, Scalability, and Cost Efficiency appeared first on MarkTechPost.

Summary

The article introduces FusionANNS, an advanced solution for approximate nearest neighbor search (ANNS), which is vital for AI applications like data mining and recommendation systems. FusionANNS enhances performance, scalability, and cost efficiency by utilizing cooperative processing between CPUs and GPUs. The technology focuses on quickly identifying the closest vectors in high-dimensional spaces, which is essential for efficiently finding similar items in various contexts.

This article was summarized using ChatGPT

Microsoft Researchers Introduce Advanced Query Categorization System to Enhance Large Language Model Accuracy and Reduce Hallucinations in Specialized Fields

Large language models (LLMs) have revolutionized the field of AI with their ability to generate human-like text and perform complex reasoning. However, despite their capabilities, LLMs need help with tasks requiring domain-specific knowledge, especially in healthcare, law, and finance. When trained on large datasets, these models often miss critical information from specialized domains, leading to […]

The post Microsoft Researchers Introduce Advanced Query Categorization System to Enhance Large Language Model Accuracy and Reduce Hallucinations in Specialized Fields appeared first on MarkTechPost.

Summary

Microsoft researchers have developed an advanced query categorization system aimed at improving the accuracy of large language models (LLMs) and reducing hallucinations, particularly in specialized fields like healthcare, law, and finance. While LLMs are capable of generating human-like text and performing complex reasoning, they struggle with domain-specific knowledge, often overlooking crucial information when trained on large datasets. This new system is designed to address these shortcomings and enhance the reliability of LLMs in specialized applications.

This article was summarized using ChatGPT