News & Updates
The latest news and updates from companies in the WLTH portfolio.
Anthropic researcher's exit sparks AI safety debate and concerns over superintelligence risks
Coxon joined Anthropic after spending three years at OpenAI working on pre-training AI models. His resignation has drawn support from a current Anthropic safety researcher, while speculation is rife online that the episode could be linked to efforts to push tighter AI regulation or keep Anthropic in the spotlight ahead of a potential initial public offering (IPO). Tanya Pandey breaks down the controversy. Anthropic researcher Jacob Coxon quitting the company has sparked a larger debate over the risks of advanced artificial intelligence (AI), how fast frontier AI labs are developing the technology and whether they can keep increasingly powerful systems under human control.Coxon joined Anthropic after spending three years at OpenAI working on pre-training AI models. His resignation has drawn support from a current Anthropic safety researcher, while
Claude Code users to get 17% less weekly usage from September 14 as Anthropic ends 50% boost
Anthropic will permanently increase standard weekly limits for Claude Code by 25%. This change follows a temporary 50% usage boost which ends September 14th. Paid users will experience a 17% reduction in their weekly Claude Code capacity. The company plans future improvements for better usage visibility and control. Anthropic is cutting the amount of Claude Code that paid users can access each week, even as the company describes the change as a permanent increase in usage limits. Starting September 14, Anthropic will permanently raise the standard weekly limits for Claude Code by 25% for users on its Pro, Max, Team and seat-based Enterprise plans. However, that will be lower than the temporary 50% increase currently available to users. In practical terms, developers will see their weekly Claude Code capacity fall by about 17% from current levels. Also Read: Salesforce, Anthropic deepen AI partnership to ramp up Claude use The company confirmed the change in posts on X on August 29. Anthropic initially said it was "permanently raising" standard weekly limits by 25%, while noting that the current 50% increase would remain in place until September 14. In a follow-up post, the company acknowledged the other side of the calculation: "Compared to today, this works out to a 17% reduction in weekly limits on Claude Code." The difference comes down to what Anthropic is using as the baseline. For example, if a user's original weekly allowance was 100 units, the temporary 50% boost increased that to 150 units. From September 14, the permanent 25% increase would take the allowance to 125 units. That means users would still have 25% more capacity than under the original limit, but 17% less than they have today. The rollback follows a period of significantly higher usage allowances for Claude Code. On May 6, Anthropic said it had agreed to a partnership with SpaceX that would substantially increase its compute capacity. The company said the SpaceX deal, along with other recent compute agreements, had allowed it to increase usage limits for Claude Code and the Claude API. As part of those changes, Anthropic said it was doubling Claude Code's five-hour rate limits for Pro, Max, Team and seat-based Enterprise plans. It also removed the peak-hours limit reduction for Pro and Max accounts and substantially increased API rate limits for Claude Opus models. Also Read: OpenAI launches ChatGPT Ads in India, over 50 brands to go live this week Anthropic subsequently introduced a temporary 50% increase to weekly Claude Code usage allowances. The boost was extended several times over the summer, making the higher allowance the effective level of access for many heavy users. Now, that extra capacity is being scaled back. The change has drawn criticism from some developers on X, with users pointing out that describing the move as a 25% increase does not capture the immediate impact on people currently using the service. "You're reducing limits even more? I used to be able to do all my work with a single MAX sub and never hit my limits. These days I run my full weekly limits in like 2 days, ur models are absolute token hogs," one user commented on Anthropic's post. Anthropic has also hinted that changes to Claude Code are coming that could make the new limits easier to manage. The company said it is working on improvements that will give users more visibility and control over their usage. "Compared to today, this works out to a 17% reduction in weekly limits," Anthropic said, adding that it was working on changes that would make it feel like users were "getting more from Claude" while giving them greater visibility and control over usage. "This is probably why I cancelled my claude subscription. At least reduce prices or do something to offset this," another user complained. The current 50% boost will remain available through September 13, before the new limits take effect on September 14.
Anthropic hires Google chip veteran Amir Salek as part of push into hardware
AI firm Anthropic has hired a Google chip founder for its new semiconductor team. This move signals Anthropic's ambition to develop its own custom AI chips. The company is also securing significant data center capacity deals. Rival OpenAI is pursuing similar custom chip development strategies. These efforts aim to address supply shortages and optimize AI hardware. Anthropic PBC has hired Amir Salek, a founder of the custom chip program at Alphabet Inc's Google, as the AI lab lays the groundwork for a push into making its own semiconductors. Salek is joining the compute team of Anthropic, maker of the Claude platform, the AI company said Friday. The executive ran Google's tensor processing unit business until 2022 and delivered the first seven generations of those chips. In the new role, Salek will report to James Bradbury. Also Read: Samsung Electronics expects shareholder returns up to $80 billion this year Anthropic, which buys chips from a variety of sources, including Nvidia Corp, Google and Amazon.com, has indicated recently that it wants to build an in-house silicon business. The San Francisco-based company has begun hiring and listing jobs for the effort. Like other AI giants, Anthropic is scrambling to get enough data center infrastructure to support its ambitions. Custom chips would potentially help it cope with supply shortages and tailor designs to better meet its needs. Also Read: South Korea's HD Hyundai Heavy says mulling US shipyard investment Rival OpenAI, the company behind ChatGPT, is taking similar steps. It has unveiled a chip called Jalapeno, co-developed with Broadcom, and plans to begin using it later this year. Anthropic is signing chip and data center capacity deals at a rapid clip - working with both established and emerging vendors. The company has an agreement with UK-based chip startup Fractile for an initial order of roughly $250 million of its chips, with the intention to expand that contract in the future, Bloomberg News has reported. The AI lab also recently signed capacity deals with Riot Platforms and Volta Infra Holdings. After leaving Google, Salek has been senior managing director at Cerberus Capital Management, a private equity firm co-founded by Deputy Secretary of Defense Stephen Feinberg. He also previously worked at Nvidia, the leading maker of AI processors.
