|
|
|
Mar 04, 2026
|
|
|
|
|
Supported by
|
|
|
|
|
|
|
Happy Wednesday! OpenAI is developing an internal alternative to Microsoft's GitHub. Apple launches a new MacBook Pro that can run AI models faster. Anthropic CEO credits the company's mission for staff retention as it hits $19 billion in annualized revenue.
|
|
|
|
OpenAI is developing an internal alternative to GitHub, Microsoft’s popular code repository that allows software engineers to store, share and collaborate on computer code, The Information reported on Tuesday. The move comes after OpenAI engineers experienced an increasing number of GitHub outages in recent months. The project is nascent and won’t be completed for many months, but staffers working on it have discussed potentially selling the code repository product to customers once it’s finished. If OpenAI does so, it
would be the latest example of the ChatGPT maker competing head-on with Microsoft, one of its largest shareholders. In the past, OpenAI has developed ChatGPT features that allow people to collaborate on documents and create and edit presentations much like people do with Microsoft Office apps, for example.
|
|
|
|
Apple announced a new MacBook Pro powered by the M5 Pro and M5 Max chipsets. The company said that the new chips inside the computers will run artificial intelligence models four times faster than the previous generation of chips. Apple also announced on Tuesday a new MacBook Air with the M5 chip and a new lineup of external computer monitors, called the Studio Display. The announcements are a part of a slew of hardware products Apple is introducing this week. On Monday, Apple announced a new low-priced iPhone and iPad models. Apple is still expected to unveil a budget-friendly MacBook this week.
|
|
|
|
Anthropic CEO Dario Amodei on Tuesday touted the company’s ability to prevent competitors from hiring away its staff and stressed the company’s interest in expanding areas like healthcare through partnerships. The maker of the Claude chatbot recently topped $19 billion in annualized revenue, or monthly revenue multiplied by 12, up from $9 billion at the end of last year, according to a person with knowledge of its finances. Amodei told attendees at an investment conference held by Morgan Stanley in San Francisco that Anthropic generally counters offers from rivals such as Meta Platforms, but wouldn’t spend 10 times employee’s current compensation just to retain them, which he described as “destructive to the culture.” Instead, the company reminds its staff they’re working at the company for the
“mission.” Amodei said that last summer, when Meta was aggressively recruiting AI researchers from OpenAI and other labs, only two of Anthropic’s employees had resigned to work at Meta, a better retention rate than for rival OpenAI, he said, according to a recording of the comments obtained by The Information. Anthropic was founded five years ago by Amodei and other staff from OpenAI after a disagreement with the ChatGPT maker over safety, or ensuring AI models don’t make serious mistakes or take actions to hurt humans. Amodei spent much of the talk describing the company’s approach to healthcare, running its AI models more efficiently and company culture. For instance, he said the company sought out partnerships in regulated industries like healthcare to expand distribution, or customer reach.
Healthcare, medicine and biology are “where the benefits of AI” are really going to come, he said. He declined to answer questions about the company’s recent dispute with the Pentagon over Anthropic’s requirement that the military avoid using its AI models for mass domestic surveillance or fully autonomous lethal weapons. Bloomberg earlier reported the revenue rate.
|
|
|
|
A key architect of Alibaba Group’s highly regarded Qwen open-source AI models said Wednesday that he is stepping down from his role, in a seemingly abrupt departure that surprised many in the industry. Junyang Lin, the Qwen team’s tech lead, said in a post on the X social media site: “Me stepping down. Bye my beloved Qwen.” Hours later, Binyuan Hui, a researcher on the Qwen team, said on X that he is also leaving, in reply to Lin’s earlier post. Under Lin’s leadership in model development, Alibaba became a major global competitor in AI, especially in the fast-growing market for open-source models. Just earlier this week, Alibaba launched small-size versions of its latest-generation Qwen3.5 AI models that can run on smartphones for on-device tasks. When Elon Musk on Tuesday wrote a post on X praising those small Alibaba models for their “impressive intelligence density,” Lin replied saying “Thx Elon!” It is unclear why Lin departed. Chen Cheng, an Alibaba employee who works closely with the Qwen team, wrote on X in reply to Lin’s post: “I’m truly heartbroken. I know
leaving wasn’t your choice. Just last night, we were side by side launching the Qwen3.5 small model. I honestly can’t imagine Qwen without you.” An Alibaba spokesperson didn’t immediately comment. Lin hasn’t replied to The Information’s request for comment sent via text messages on X.
|
|
|
|
OpenAI is considering an agreement to deploy its AI on all North Atlantic Treaty Organization unclassified networks, an OpenAI spokesperson said, days after OpenAI announced it would start supplying the Defense Department with AI on its classified networks. CEO Sam Altman made the comment about NATO during an employee meeting in which he defended the company from criticism, including from some OpenAI employees, that the company’s recent agreement with the Defense Department would allow the military to use its AI in violation of the company’s prohibitions. The OpenAI spokesperson declined to share any additional details about the NATO contract or the status of discussions about it. OpenAI has previously said it is open to classified contracts with countries in alliances with the United States. On Monday, OpenAI said that it had updated its Defense Department agreement with more specific language preventing the use of its AI for domestic surveillance, including via commercially acquired information. Altman is expected to speak on Thursday at Morgan Stanley’s Tech, Media and Telecom conference in San Francisco. The Wall Street Journal first reported details of the meeting. (After this story published, an OpenAI spokesperson clarified that the company was considering a contract for NATO unclassified networks, not classified networks. It has been updated.)
|
|
|
|
| |
