|
|
|
Apr 09, 2026
|
|
|
|
|
Supported by
|
|
|
|
|
|
|
Happy Thursday! Meta Platforms unveils new AI models. Former Andreessen Horowitz general partner Anjney Midha raises $1.3 billion. A federal appeals court rejects Anthropic’s motion to halt the Pentagon’s blacklisting of the company.
|
|
|
|
Meta Platforms CEO Mark Zuckerberg has announced the availability of a new family of AI models on Wednesday. In a Threads post, Zuckerberg said the company is releasing its first model, Spark, as part of a broader suite of models called Muse. Spark will power Meta AI, the company’s internally developed AI assistant. He added that it is “particularly strong in areas related to personal superintelligence,” including visual understanding, health, social content, shopping, games, and more. The Meta chief also said the parent company of Facebook, Instagram and Threads plans to release agent-based products alongside increasingly advanced models in the future, including open-source releases. Spark is the first AI model to emerge from the company’s AI division, Meta Superintelligence Labs, which was formed nine months ago. Last year, Meta faced challenges with the performance of its large language model, Llama 4, prompting it to delay the model’s release. When Meta eventually launched two versions of Llama 4, named Maverick and Scout, some developers expressed disappointment with the software’s performance.
|
|
|
|
Former Andreessen Horowitz general partner Anjney Midha has secured $1.3 billion for the first fund for AMP, his newly formed venture firm, according to a source with knowledge of the raise. AMP recently made a $300 million investment in Anthropic as part of the Claude maker’s $30 billion round that valued it at $380 billion including the investment. Midha had been one of the first personal investors in Anthropic. Last year Midha departed his role at Andreessen Horowitz, where he invested in AI model
maker Mistral and image maker Black Forest Labs. He remains a venture partner at the firm. AMP has already started raising capital for its second venture fund. Midha has also been working on a $10 billion infrastructure credit fund. Midha has told The Information that AMP, is creating a “grid for AI,” similar to how electricity access is managed by a central grid, that will reshape how AI developers access hard-to-get servers.
|
|
|
|
A federal appeals court on Wednesday rejected Anthropic’s motion to halt the Department of Defense’s blacklisting of the company while Anthropic challenges the designation in court. The ruling from the circuit court in Washington, D.C., sets up split decisions in Anthropic’s two lawsuits challenging the Pentagon’s designation of it as a supply chain risk. Each suit challenges a separate statute used by the Pentagon. Last month, a federal judge in San Francisco approved a preliminary injunction, which went into effect last week, halting the supply chain risk designation under one statute. The new order isn’t a final decision in the D.C. case. The circuit court set an expedited schedule to hear that lawsuit, with oral arguments scheduled for May 19. In practice, the D.C. court’s decision means that the Pentagon can still treat Anthropic as a supply chain risk while the lawsuits proceed, excluding it from new military contracts. However, the San Francisco injunction still blocks President Donald Trump’s order banning Anthropic in use from other federal agencies. In explaining their decision, the three judges in the D.C. case wrote: “On one side is a relatively contained risk of financial harm to a single private company. On the other side is judicial management of how, and through whom, the Department of War secures vital AI technology during an active military conflict.” Charlie Bullock, a senior research fellow at the Institute for Law and AI who has been following the case, said that the decision wasn’t surprising. “2/3 judges on the D.C. Circuit panel
have been very, very sympathetic to the Trump administration’s aggressive claims about executive authority in the past,” he said.
|
|
|
|
Meta Platforms has taken down an internal, employee-built leaderboard tracking how many tokens staffers were using. The dashboard, which employees referred to as Claudeonomics, showed total usage over a recent 30-day period across the company amounted to over 60 trillion tokens, The Information reported on Monday. The top-ranked individual user averaged 281 billion tokens, which could cost in the hundreds or thousands of dollars. The dashboard now displays a message that it is offline. It says: “We’ve really enjoyed building this
app on Nest for everyone. It was meant to be a fun way for people to look at tokens, but due to data from this dashboard being shared externally, we’ve made the decision to shutter Claudeonomics for now.” (Nest is reportedly an internal software development tool at Meta.) The leaderboard tracked a new trend in Silicon Valley called tokenmaxxing, a response to employers who have made token usage a proxy for productivity and a way to compete over who is most AI-native. In Meta’s case, employees competed for a place on its leaderboard of the top 250 token users, along with badges such as “Session
Immortal” or “Token Legend.”
|
|
|
|
As investors fret that AI startups will gobble up the business of established software companies, a top executive from Workday has jumped ship for Anthropic. Peter Bailis, who joined the human resources software giant provider Workday last May as its chief technology officer, left last month and is now a member of technical staff at Anthropic, according to his LinkedIn profile. A Workday spokesperson confirmed his departure. An Anthropic spokesperson said Bailis will be working on reinforcement learning engineering that is focused on getting the startup’s AI models to run faster and more reliably. Workday has promoted Gabe Monroy, a senior vice president who joined the company last August from Google Cloud, to replace Bailis as CTO, according to the
spokesperson. Anthropic has shown interest in building its own HR offerings. It recently posted a job listing for an engineering manager to oversee all aspects of developing “people products” that help the startup hire new employees, train and develop them, and manage promotions. All of these functions are ones that Workday’s applications are designed to handle. Anthropic was reportedly a customer of Workday as recently as February. And in a January job posting for a head of IT regulatory compliance, Anthropic said it was looking for candidates with experience with Workday, as well as Salesforce and NetSuite.
|
|
|
|
| |
