Month: June 2024

Enlarge (credit: Getty Images)

On Tuesday, a group of former OpenAI and Google DeepMind employees published an open letter calling for AI companies to commit to principles allowing employees to raise concerns about AI risks without fear of retaliation. The letter, titled "A Right to Warn about Advanced Artificial Intelligence," has so far been signed by 13 individuals, including some who chose to remain anonymous due to concerns about potential repercussions.

The signatories argue that while AI has the potential to deliver benefits to humanity, it also poses serious risks that include "further entrenchment of existing inequalities, to manipulation and misinformation, to the loss of control of autonomous AI systems potentially resulting in human extinction."

They also assert that AI companies possess substantial non-public information about their systems' capabilities, limitations, and risk levels, but currently have only weak obligations to share this information with governments and none with civil society.

Enlarge

A ransomware attack that crippled a London-based medical testing and diagnostics provider has led several major hospitals in the city to declare a critical incident emergency and cancel non-emergency surgeries and pathology appointments, it was widely reported Tuesday.

The attack was detected Monday against Synnovis, a supplier of blood tests, swabs, bowel tests, and other hospital services in six London boroughs. The company said it has "affected all Synnovis IT systems, resulting in interruptions to many of our pathology services." The company gave no estimate of when its systems would be restored and provided no details about the attack or who was behind it.

Major impact

The outage has led hospitals, including Guy's and St Thomas' and King's College Hospital Trusts, to cancel operations and procedures involving blood transfusions. The cancellations include transplant surgeries, which require blood transfusions.

Enlarge / All these pieces more or less pop apart after a brief chemical treatment. (credit: Israel Sebastian.)

For years, the arguments against renewable power focused on its high costs. But as the price of wind and solar plunged, the arguments shifted. Suddenly, concerns about the waste left behind when solar panels hit end-of-life became so common that researchers at the US's National Renewable Energy Lab felt compelled to publish a commentary in Nature Physics debunking them.

Part of the misinformation is pure nonsense. The primary ingredients of most panels are silicon, aluminum, and silver, none of which is a major environmental threat. Solar panels also have a useful lifespan of decades, and the vast majority of those in existence are less than 10 years old, so waste hasn't even become much of a problem yet. And, even once these panels age out, recycling techniques are available.

Perhaps the only realistic concern is that existing recycling technologies rely on nitric acid and can produce some toxic waste. But a group of researchers from Wuhan University have figured out an alternative means of recycling that avoids the production of toxic waste, and is more energy-efficient as a bonus.

Enlarge / Some subscribers say they'll just cancel after watching Season 2 of House of the Dragon. (credit: Warner Bros. Discovery)

Max's monthly ad-free streaming subscription costs $1 more than it did yesterday. Warner Bros. Discovery (WBD) today raised the prices for ad-free Max subscription plans, effective immediately for new subscribers.

In addition to the cheaper ad-free plan now costing $17 per month, the Ultimate ad-free monthly plan, which supports four instead of two concurrent streams and 4K instead of 1080p resolution, is up a dollar to $21 per month.

Annual subscriptions to the streaming service are also getting more expensive. The annual ad-free plan is $20 more, at $170/year, and the Ultimate ad-free annual plan went up by $10 to $210 per year.

Enlarge / Pixel devices. (credit: Google)

Google has changed its repair policy in response to criticism from repair advocate Louis Rossmann. Rossmann dug through the Google Store's "Service & Repair Program Terms & Conditions" for its first-party mail-in repair service and found the same style of onerous bans on third-party parts that Samsung was recently caught using. Section D, article 4 of the terms include the rather incredible line "Unauthorized Parts: You will not send in a Device containing non-Google-authorized parts – if You do, Your Device will not be returned to you." That's right, according to the terms, Google would keep a device sent in for repair, and you wouldn't get it back.

We asked Google for a comment on Rossmann's video, and a spokesperson says the terms will be updated:

If a customer sends their Pixel to Google for repair, we would not keep it regardless of whether it has non-OEM parts or not. In certain situations, we won't be able to complete a repair if there are safety concerns. In that case, we will either send it back to the customer or work with them to determine next steps. Customers are also free to seek the repair options that work best for them. We are updating our Terms and Conditions to clarify this.

That sounds a lot more reasonable.

Enlarge (credit: Getty Images)

Zoom CEO Eric Yuan has a vision for the future of work: sending your AI-powered digital twin to attend meetings on your behalf. In an interview with The Verge's Nilay Patel published Monday, Yuan shared his plans for Zoom to become an "AI-first company," using AI to automate tasks and reduce the need for human involvement in day-to-day work.

"Let’s say the team is waiting for the CEO to make a decision or maybe some meaningful conversation, my digital twin really can represent me and also can be part of the decision making process," Yuan said in the interview. "We’re not there yet, but that’s a reason why there’s limitations in today’s LLMs."

LLMs are large language models—text-predicting AI models that power AI assistants like ChatGPT and Microsoft Copilot. They can output very convincing human-like text based on probabilities, but they are far from being able to replicate human reasoning. Still, Yuan suggests that instead of relying on a generic LLM to impersonate you, in the future, people will train custom LLMs to simulate each person.