Worth Knowing
U.S. and Chinese Firms Respond to Export Controls With Alternative Chips: Major semiconductor companies in the United States and China are adapting to the new U.S. export controls that placed strict performance limits on the chips that can be sold to Chinese customers — U.S.-based Nvidia has begun selling a less powerful version of one of its high-end datacenter GPUs, and Chinese firms are reportedly working to field their own chips that comply with the controls. Reuters reports that Nvidia’s A800 GPU is an “alternative product to the Nvidia A100 GPU for customers in China” that went into production earlier this year. According to Reuters, the new Nvidia chips are already listed for sale by some Chinese retailers. Chinese companies, meanwhile, appear further behind in rolling out their own compliant chips. As we covered recently, TSMC suspended production for Shanghai-based Biren Technology over concerns that its high-end GPUs exceeded the performance limits set by the new controls. Now the Financial Times reports that both Biren and Chinese tech giant Alibaba are working to find a compliant design without starting from scratch, which could lead to expensive delays.
Tech Layoffs Could Signal a Downturn — What Do They Mean for AI?: A wave of layoffs at a number of major tech companies has some observers concerned about the future of the industry, but signals are mixed when it comes to AI. Headlined by job cuts at Twitter and Meta — which laid off 50 and 13 percent of their respective workforces — the cuts also impacted staff at Amazon, Microsoft, Snap and a number of other firms. Some of the job losses hit teams working on AI and machine learning — Twitter laid off its entire ML Ethics, Transparency and Accountability team and Meta let go its 50-person ML-focused “Probability” team. But other AI projects appear to be safe for now. Meta’s AI lab seemingly escaped the chopping block, and major AI programs at the likes of Alphabet and Microsoft similarly appear to have remained intact. For those interested in the future of AI investment and research, there are even some glimmers of hope in the recent news — in a letter to employees, Meta CEO Mark Zuckerberg said the company would be funneling resources into “high priority growth areas,” including AI. In the face of worsening forecasts and sinking stocks, more layoffs wouldn’t come as a surprise. But so long as executives remain optimistic about the importance of AI — according to a Deloitte survey released last month, 94 percent of business leaders agreed that “AI is critical to success over the next five years” — it would be surprising to see the dramatic cuts characteristic of an “AI winter.”
- More: Why Meta and Twitter’s AI and ML Layoffs Matter | The Collapse of Sam Bankman-Fried’s FTX Riles A.I. Research Community
New High-End CPUs from AMD and Intel — Should Intel Be Worried?: AMD announced its new line of server-grade CPUs, and Intel unveiled its newest server CPUs and GPUs. While GPUs tend to get more attention in the AI space, CPUs are still critical components — both of Intel’s new chips are set to power the Aurora exascale supercomputer at Argonne National Laboratory, which is scheduled to power up next year. But despite that impressive deployment, observers say the winner here is likely AMD. Its new Epyc CPUs, based on TSMC’s 5 nm process, appear to be significantly faster than Intel’s Max Series CPUs (we’ll have to wait for both chips to be released for exact performance comparisons), which are based on Intel’s 10 nm-equivalent process node (though nanometer-based designations are more marketing than definitive measurement) and manufactured in Arizona. Intel has been steadily losing ground to AMD CPUs in the supercomputer space and the broader server market — if both chips perform as expected, it wouldn’t be a surprise to see that trend continue. That would be welcome news for AMD, but it would be less positive for U.S. high-end chip manufacturing. While TSMC’s 5 nm process is coming to Arizona in 2024 (see the entry above), for now, all Epyc CPUs will be manufactured in Taiwan.
Government Updates
DARPA Tests Autonomous Black Hawk at Project Convergence 2022: The Defense Advanced Research Projects Agency tested an autonomous Black Hawk helicopter as part of Project Convergence 2022, the Army-hosted joint exercise meant to test the military’s newest technologies. While DARPA and Sikorsky — which built the Black Hawk’s MATRIX aircraft autonomy system — successfully completed a 30-minute flight with no pilots aboard in February, the latest tests saw the Black Hawk stretch its wings (or rotors): Sikorsky parent Lockheed Martin says the helicopter performed both an 83-mile “Long-endurance Medical Resupply” mission, during which it carried 500 pounds of real and simulated blood, and a combined “Cargo Delivery and Casualty Evacuation” mission, in which it flew for 30 minutes with a 2,600 pound load suspended by a 40-foot sling, delivered the cargo after a planned redirection, landed to evacuate a simulated casualty, and returned home (the company posted a video of both tests on its YouTube page). The tests are part of DARPA’s Aircrew Labor In-Cockpit Automation System (ALIAS) program, which aims to add automation to existing aircraft. DOD officials say that in addition to enabling unmanned operation for potentially dangerous missions, ALIAS could also help to ease pilot workloads by making single-pilot or pilotless flights possible for more routine operations.
Despite U.S. Push, New Multilateral Export Controls Remain Uncertain: U.S. officials are working hard to get allies to join them in imposing strict new export controls on China’s semiconductor industry, but so far, an agreement remains elusive. As CSET Senior Fellow Kevin Wolf noted after the United States announced its sweeping package of restrictions last month, the controls stand a much better chance of working if the United States can convince other major semiconductor manufacturing equipment (SME) producers — especially the Netherlands and Japan — to put up a united front. In public comments late last month, Under Secretary of Commerce for Industry and Security Alan Estevez struck an optimistic note, saying he expected a deal soon. But as the Financial Times reports, that optimism is less pronounced in the Hague, where officials have been frustrated by the aggressive U.S. approach. Bloomberg echoed those rumors and reported a similar level of annoyance in Tokyo. Estevez and National Security Council Senior Director for Technology and National Security (and former CSET Senior Fellow) Tarun Chhabra are traveling to the Netherlands this month for further discussions, and Politico reports that semiconductor controls will also be on the docket when U.S. and EU officials gather next month for the third U.S.-EU Trade and Technology Council meeting. Neither is expected to yield a concrete agreement, however, and Commerce Secretary Gina Raimondo has reportedly warned U.S. SME manufacturers that it could be up to nine months before any such agreement is reached.
In Translation
CSET’s translations of significant foreign language documents on AI
CSET’s translations of significant foreign language documents on AI
PRC Ministry of Science and Technology Application Guide: Application Guide for the 2021 First Batch of Projects for the “Intergovernmental International Cooperation on Science and Technology Innovation” Key Special Project (Draft for Feedback). This guide, written by China’s Ministry of Science and Technology, lists international S&T cooperation projects that Chinese companies and research institutes can apply for participation in. Participating firms and institutes receive funding from the PRC government. The projects involve a wide range of countries and international organizations and cover a vast array of different technologies.
If you have a foreign-language document related to security and emerging technologies that you’d like translated into English, CSET may be able to help! Click here for details.
CSET Job Openings
We’re hiring! Please apply or share the roles below with candidates in your network:
- Fellow – Emerging Technology Supply Chains: We are currently seeking candidates to lead and coordinate our Emerging Technology Supply Chains Line of Research, either as a Research Fellow or Senior Fellow (depending on experience). This fellow will shape priorities, lay out an overall research strategy, oversee execution of the research and production of reports, and help hire and manage supporting researchers. Applications due by December 12.
What’s New at CSET
REPORTS
- China’s AI Workforce: Assessing Demand for AI Talent by Diana Gehlhaus, Joanne Boisson, Sara Abdulla, Jacob Feldgoise, Luke Koslosky and Dahlia Peterson
- A Plea: The Case for Digital Environmentalism by Andrew Burt and Daniel E. Geer, Jr.
- CSET: Data Snapshot: Global GitHub: Mapping the Utilization of Open Source Software by Organizations by Christian Schoeberl
- Nature: How Scientists Can Inform Policy Decisions by Dewey Murdick
- ChinaTalk Podcast: Export Controls for AI: Will They Work? with Emily Weinstein and Tim Hwang
- On November 10, CSET Research Analysts Dahlia Peterson and Luke Koslosky discussed Chinese AI workforce trends with AMPLYFI’s Head of Professional Services, Oliver Hayman.
- Axios: Jennifer A. Kingson cited the 2021 report Truth, Lies, and Automation: How Language Models Could Change Disinformation in an article about Google’s AI developments.
- NextGov: An Edward Graham piece about the U.S.-China Economic and Security Review Commission’s report to Congress noted the findings of Jack Corrigan, Sergio Fontanez and Michael Kratsios’ recent brief, Banned in D.C.: Examining Government Approaches to Foreign Technology Threats.
- The Register: The latest brief by Diana Gehlhaus, Joanne Boisson, Sara Abdulla, Jacob Feldgoise, Luke Koslosky and Dahlia Peterson — China’s AI Workforce: Assessing Demand for AI Talent — earned a writeup by Laura Dobberstein.
- ChinAI: Jeffrey Ding named China’s AI Workforce a “Should-read” piece in the most recent edition of his ChinAI newsletter.
What We’re Reading
Report: 2022 Annual Report to Congress, United States-China Economic and Security Review Commission (November 2022)
Article: Mending the “Broken Arrow”: Confidence Building Measures at the AI-Nuclear Nexus, Lauren Kahn, War on the Rocks (November 2022)
Upcoming Events
- December 1-2: The Athens Roundtable: Artificial Intelligence and the Rule of Law
- December 5: CSET Webinar, Introducing the Emerging Technology Observatory, featuring Zachary Arnold, Melissa Flagg and John VerWey
What else is going on? Suggest stories, documents to translate & upcoming events here.