Tone'o Down-Under has added a photo to the pool:
Trinity Church. Part of old Perth that has not yet been torn down to make way for glass and concrete.
On St Georges Terrace, Perth, Western Australia.
/s3/static.nrc.nl/images/gn4/stripped/data137237752-dc3db8.jpg)
Microsoft and OpenAI have signed a non-binding deal to restructure their partnership, paving the way for OpenAI to shift into a conventional for-profit model and potentially go public. Reuters reports: Details on the new commercial arrangements were not disclosed, but the companies said they were working to finalize terms of a definitive agreement. [...] Microsoft invested $1 billion in OpenAI in 2019 and another $10 billion at the beginning of 2023. Under their previous agreement, Microsoft had exclusive rights to sell OpenAI's software tools through its Azure cloud computing platform and had preferred access to the startup's technology.
Microsoft was once designated as OpenAI's sole compute provider, though it lessened its grip this year to allow OpenAI to pursue its own data center project, Stargate, including signing $300 billion worth of long-term contracts with Oracle, as well as another cloud deal with Google. As OpenAI's revenue grows into the billions, it is seeking a more conventional corporate structure and partnerships with additional cloud providers to expand sales and secure the computing capacity needed to meet demand. Microsoft, meanwhile, wants continued access to OpenAI's technology even if OpenAI declares its models have reached humanlike intelligence - a milestone that would end the current partnership under existing terms.
OpenAI said under current terms, its nonprofit arm will receive more than $100 billion -- about 20% of the $500 billion valuation it is seeking in private markets -- making it one of the most well-funded nonprofits, according to a memo from Bret Taylor, chairman of OpenAI's current nonprofit board. The companies did not disclose how much of OpenAI Microsoft will own, nor whether Microsoft will retain exclusive access to OpenAI's latest models and technology. Regulatory hurdles remain for OpenAI, as attorneys general in California and Delaware need to approve OpenAI's new structure. The company hopes to complete the conversion by year's end, or risk losing billions in funding tied to that timeline.
Microsoft was once designated as OpenAI's sole compute provider, though it lessened its grip this year to allow OpenAI to pursue its own data center project, Stargate, including signing $300 billion worth of long-term contracts with Oracle, as well as another cloud deal with Google. As OpenAI's revenue grows into the billions, it is seeking a more conventional corporate structure and partnerships with additional cloud providers to expand sales and secure the computing capacity needed to meet demand. Microsoft, meanwhile, wants continued access to OpenAI's technology even if OpenAI declares its models have reached humanlike intelligence - a milestone that would end the current partnership under existing terms.
OpenAI said under current terms, its nonprofit arm will receive more than $100 billion -- about 20% of the $500 billion valuation it is seeking in private markets -- making it one of the most well-funded nonprofits, according to a memo from Bret Taylor, chairman of OpenAI's current nonprofit board. The companies did not disclose how much of OpenAI Microsoft will own, nor whether Microsoft will retain exclusive access to OpenAI's latest models and technology. Regulatory hurdles remain for OpenAI, as attorneys general in California and Delaware need to approve OpenAI's new structure. The company hopes to complete the conversion by year's end, or risk losing billions in funding tied to that timeline.
Read more of this story at Slashdot.
Intel's long-time Xeon chief architect Ronak Singhal is leaving the company after nearly 30 years, marking yet another high-profile departure amid Intel's leadership churn and intensifying competition from AMD and Arm-based cloud CPUs. The Register reports: The Carnegie Mellon alum holds degrees in electrical and computer engineering, along with at least 30 patents involving CPUs. Singhal joined Intel in 1997 after spending the previous summer as an intern at Cyrix. After a year in Intel's Rotation Engineers Program, he spent the remainder of his tenure helping to develop some of the chipmaker's most consequential and, at times, controversial processors. Most notably, Singhal oversaw the core development of Intel's 22nm Haswell and 14nm Broadwell processor architectures. His innovations aren't limited to the datacenter either, with his architectural contributions playing a significant role in the success of Intel's Core and Atom processor families as well. [...]
Singhal is only the latest Xeon lead to jump ship since the start of the year. In January, Sailesh Kottapalli, another senior fellow, left for Qualcomm barely a month after former CEO Pat Gelsinger's unceremonious "retirement." Even before Gelsinger's eviction, Intel's datacenter group has been something of a revolving door. Last summer Singhal's long-time colleague Lisa Spelman departed the company, eventually landing a spot as CEO of HPC interconnect vendor Cornelis Networks. Her replacement, Ryan Tabrah, lasted seven months in the role, about half as long as Intel datacenter boss Justin Hotard, who defected for the forests of Finland to lead Nokia as its new President and CEO back in April.
In fact, the churn now extends all the way to the top. On Monday, Intel announced its CEO of Products, Michelle Johnston Holthaus, would be leaving the business. The move is part of a broader executive shakeup that will see former Arm engineer Kevork Kechichian take over as head of Intel's datacenter engineering group. Jim Johnson, meanwhile, will take over as head of the chipmaker's client computing group while Srinivasan (Srini) Iyengar will head up a new central engineering division.
Singhal is only the latest Xeon lead to jump ship since the start of the year. In January, Sailesh Kottapalli, another senior fellow, left for Qualcomm barely a month after former CEO Pat Gelsinger's unceremonious "retirement." Even before Gelsinger's eviction, Intel's datacenter group has been something of a revolving door. Last summer Singhal's long-time colleague Lisa Spelman departed the company, eventually landing a spot as CEO of HPC interconnect vendor Cornelis Networks. Her replacement, Ryan Tabrah, lasted seven months in the role, about half as long as Intel datacenter boss Justin Hotard, who defected for the forests of Finland to lead Nokia as its new President and CEO back in April.
In fact, the churn now extends all the way to the top. On Monday, Intel announced its CEO of Products, Michelle Johnston Holthaus, would be leaving the business. The move is part of a broader executive shakeup that will see former Arm engineer Kevork Kechichian take over as head of Intel's datacenter engineering group. Jim Johnson, meanwhile, will take over as head of the chipmaker's client computing group while Srinivasan (Srini) Iyengar will head up a new central engineering division.
Read more of this story at Slashdot.