Time for a catch-up!

This week seems to have been a lot quieter than usual, perhaps because everyone has survived the summer, and are now simply trying to settle back into work mode.

In any event, it’s been a busy few weeks for all of those human resource experts and employment lawyers in the World of AI, as no sooner had Meta snapped up droves of experts for the Super Intelligence Labs project, than almost a dozen left again, some only having stayed for days and one said to have never even showed up for work. All of this despite compensation packages running into the tens of millions of dollars. Meta instituted a hiring freeze and a re-structuring.

It's not just Meta that is facing this employment upheaval, as xAI is also undergoing significant churn, particularly amongst senior players. This week saw the departure after only 5 months of Mike Liberatore, the CFO, who had only joined in April after leaving Airbnb. This follows the resignations of xAI’s chief legal advisor, other senior lawyers, and co-founder Igor Babuschkin, and it also comes after CEO Linda Yaccarino stepped down following concerns about xAI’s chatbot Grok.

The series of departures raises concerns about stability inside xAI, especially as the company races to fundraise and expand Grok. Leadership turbulence could impact investor confidence and slow down its ambitious push to compete in the AI space.

Anyway, let’s dive right in on the other news from last week …

Big Sharks vs. small fish

Unusually, there were no major software releases or updates, but the big news was OpenAI buying Statsig, a Seattle-based startup, for $1.1 billion in stock. This is an interesting move by OpenAI.

Statsig has helped companies like Microsoft, Notion, SoundCloud and OpenAI, to test various features on their new products. This experimentation is critical for ensuring that OpenAI is able to ensure the apps and products it produces are what people actually want, using a technic called A/B testing. This again underlines an important, if subtle shift in OpenAI’s focus.

In China, Alibaba is reportedly developing its own AI chip, triggering nearly a 13% surge in its stock. In contrast, shares of NVIDIA, AMD, and others slumped on concerns over potential reduced AI chip demand from China. 

Nvidia were busy again this week, even after their very positive sales figures a week ago. First of all, Jensen Huang stated that he sees AI driving a shift toward a four-day working week, citing increased productivity and economic growth. However, he added that it doesn’t necessarily mean less work, but rather that workers might actually be busier, albeit more efficient. 

Then they announced about the release of a chip called Thor. This is a tiny chip that fits in the palm of your hand but thinks like a data centre running up to 7 AI models simultaneously, providing 7.5 more computing power whilst using 3.5 less energy. This is significant, as it is already being used by Amazon in their robots, and as Huang pointed out, their robotics division grew by 72% last year.

Legislation, policy and other news

The big news policy and legislation wise was Thursday’s “launch” event for the Presidential AI Challenge, hosted by the First Lady Melania Trump at the White House. She has managed to get 135 companies to pledge billions to teach kids in the US about artificial intelligence. Google's CEO Sundar Pichai showed up and dropped $150 million on the table. IBM promised to train 2 million Americans in AI skills, and Microsoft threw in $1.25 million just for prizes in a student AI competition.

In the evening the President then hosted a gala dinner for the key people in the AI industry, although most of the commentary revolved around the fact that Elon Musk was not invited.

Moving onto some positive news, there were press releases on a number of major medical breakthroughs, supported by AI this week. The first was news that Researchers at Imperial College London, in collaboration with Eko Health, unveiled a cutting-edge AI-powered stethoscope capable of diagnosing heart failure, valve disease, and atrial fibrillation within just 15 seconds.

In UK trials involving around 12,000 symptomatic patients, the device demonstrated significantly higher detection rates - doubling for heart failure, tripling for atrial fibrillation, and nearly doubling for valve disease. The innovation promises to transform primary care diagnostics, though experts advise cautious use to avoid false positives. 

The UK NHS then announced that it has has deployed a breakthrough AI tool across all 107 stroke centres in England. It accelerates CT brain scan analysis, cutting treatment time from 140 to 79 minutes, and boosts patient recovery rates from 16% to 48%.

Finally, on the other side of the Atlantic, AI assisted processes allowed a company named Recursion to move candidates to clinical trials after 14 months, compared with the previous industry “average” time of 42 months.

That’s all for now from me at the end of yet another week … so stay informed, stay critical, and wherever possible - stay ahead.

Regards

Tom Carter