tech

Jon Keegan4/1/25

Report: Google’s DeepMind is withholding AI research for competitive advantage

Last year, GoogleDeepMind CEO Demis Hassabis won the Nobel Prize for chemistry (along with his colleague John M. Jumper) for his work on AlphaFold2, a groundbreaking model that’s been used by over 2 million researchers in 190 countries.

Now, the Financial Times is reporting that former researchers who worked with Hassabis at DeepMind say the organization is breaking with that tradition of scientific sharing by withholding AI research for Google’s strategic advantage in a very competitive market.

According to the report, researchers said work that showed weaker performance by Google’s own models was held back, as well as research that highlighted weaknesses in competitor OpenAI’s models.

Most of today’s modern AI tools have Google research to thank for their existence. The key innovation of “transformers” was developed at Google, but the company was late to capitalize on the breakthrough.

DeepMind slows down research releases to keep competitive edge in AI race

DeepMind slows down research releases to keep competitive edge in AI race

According to the report, researchers said work that showed weaker performance by Google’s own models was held back, as well as research that highlighted weaknesses in competitor OpenAI’s models.

Most of today’s modern AI tools have Google research to thank for their existence. The key innovation of “transformers” was developed at Google, but the company was late to capitalize on the breakthrough.

More Tech

Salesforce is using AI to to handle customer service, and it’s saving the company $100 million per year, CEO Marc Benioff said at the company’s Dreamforce conference, per Bloomberg reporting. Benioff also announced that 12,000 customers are using its “Agentforce” AI-driven customer service platform.

$100 million seems impressive, but to put that number in perspective, last quarter, the company reported over $10 billion in revenue.

Benioff has enthusiastically embraced the use of AI and has slashed thousands of positions as the company automates roles.

tech

Sam Altman says OpenAI fixed ChatGPT’s serious mental health issues in just a month. Anyway, here comes the erotica

Well that was quick. Just over a month ago, OpenAI CEO Sam Altman announced a 120-day plan to roll out new protections for identifying and helping ChatGPT users who are suffering a mental health crisis, after a series of reports brought attention to such users harming themselves and others after using the company’s AI chatbot.

Today, Altman says the company has built new tools to address these issues and “mitigated” these problems.

Altman is so confident that they’ve addressed mental health safety that the company is reverting ChatGPT’s behavior so it “behaves more like what people liked about 4o.” Altman essentially apologized to users for the changes that were made to address mental health problems that arose with use of the chatbot:

“We realize this made it less useful/enjoyable to many users who had no mental health problems, but given the seriousness of the issue we wanted to get this right.”

Separately, the company announced the members of its Expert Council on Well-Being and AI, an eight-person council of mental health experts.

As a reward for the adults who aren’t suffering mental health issues exacerbated by confiding in the chatbot, Altman says that erotica is on the way.

“In December, as we roll out age-gating more fully and as part of our ‘treat adult users like adults’ principle, we will allow even more, like erotica for verified adults.”

In response to Altman’s post on X, Missouri Senator Josh Hawley quoted Altman’s post with this message:

“You made ChatGPT ‘pretty restrictive’? Really. Is that why it has been recommending kids harm and kill themselves?”

Sam Altman on X: "We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. " / X

Sam Altman on X: "We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. We realize this made it less useful/enjoyable to many users who had no mental health problems, but given the seriousness of the issue we wanted to get this right. Now that we have" / X

Altman is so confident that they’ve addressed mental health safety that the company is reverting ChatGPT’s behavior so it “behaves more like what people liked about 4o.” Altman essentially apologized to users for the changes that were made to address mental health problems that arose with use of the chatbot:

“We realize this made it less useful/enjoyable to many users who had no mental health problems, but given the seriousness of the issue we wanted to get this right.”

Separately, the company announced the members of its Expert Council on Well-Being and AI, an eight-person council of mental health experts.

As a reward for the adults who aren’t suffering mental health issues exacerbated by confiding in the chatbot, Altman says that erotica is on the way.

“In December, as we roll out age-gating more fully and as part of our ‘treat adult users like adults’ principle, we will allow even more, like erotica for verified adults.”

In response to Altman’s post on X, Missouri Senator Josh Hawley quoted Altman’s post with this message:

“You made ChatGPT ‘pretty restrictive’? Really. Is that why it has been recommending kids harm and kill themselves?”

tech

Meta says Instagram teen accounts will default to a PG-13 content limit

Meta is introducing new guidelines for the content on Instagram teen accounts. The company is turning to the well-known PG-13 standard from the Motion Picture Association, used by the film industry.

Any user under the age of 18 will have their content limited to PG-13.

Parents who administer their child’s teen account will have the ability to change the settings — including placing their child in a more restrictive level than PG-13 — but that assumes the teen hasn’t just tried to sign up on their own using a fake birthday.

To counter those wily kids, Instagram will use “age prediction technology” to set content restrictions, according to the company.

In a blog post announcing the new policy, Meta acknowledged the new settings may not catch all prohibited content:

“Just like you might see some suggestive content or hear some strong language in a PG-13 movie, teens may occasionally see something like that on Instagram — but we’re going to keep doing all we can to keep those instances as rare as possible.”

New PG-13 Guidelines for Instagram Teen Accounts | Meta

New PG-13 Guidelines for Instagram Teen Accounts | Meta

Parents who administer their child’s teen account will have the ability to change the settings — including placing their child in a more restrictive level than PG-13 — but that assumes the teen hasn’t just tried to sign up on their own using a fake birthday.

To counter those wily kids, Instagram will use “age prediction technology” to set content restrictions, according to the company.

In a blog post announcing the new policy, Meta acknowledged the new settings may not catch all prohibited content:

“Just like you might see some suggestive content or hear some strong language in a PG-13 movie, teens may occasionally see something like that on Instagram — but we’re going to keep doing all we can to keep those instances as rare as possible.”

tech

Smartphone upgrades grew for Apple and Samsung last quarter

The global smartphone market grew 2.6% in the third quarter, thanks in part to interest in the latest phones from Apple and Samsung, according to new shipment data from market intelligence firm IDC.

“Apple and Samsung posted strong results as their latest devices encouraged consumers to upgrade in the premium segment, while new, affordable AI-enabled smartphones also drove high upgrades in more affordable price categories,” IDC Vice President of Client Devices Francisco Jeronimo said in a press release for the data, which would include roughly half a month of new iPhone sales. “Demand for Apple’s new iPhone 17 lineup was robust, with pre-orders surpassing those of the previous generation. At the same time, Samsung’s Galaxy Z Fold 7 and Galaxy Z Flip 7 outperformed all earlier foldable models, creating renewed momentum for the foldables segment.”

Here’s the year-over-year growth in third-quarter shipments:

And here’s how the absolute number of shipments compared last quarter:

The “other” bin is made up of dozens of smaller, often regional and low-cost manufacturers.

tech

Jon Keegan10/13/25

Sora’s ghoulish reanimation of dead celebrities raises alarms

OpenAI’s video generation app Sora has spent its first two weeks at the top of the charts.

The startup’s fast-and-loose approach to enforcing intellectual property rights has seen the app flooded with videos of trademarked characters in all sorts of ugly scenarios.

But another area where Sora users have been pushing the limits involves videos that reanimate dead celebrities.

And we’re not talking just JFK, MLK, and Einstein. Videos featuring more recently deceased figures such as Robin Williams (11 years ago), painter Bob Ross (30 years ago), Stephen Hawking (seven years ago), and even Queen Elizabeth II (three years ago) have been generated. Some of the videos are racist and offensive, shocking the relatives of the figures.

OpenAI told The Washington Post that it is now allowing representatives of “recently deceased” celebrities and public figures to request that their likenesses be blocked from the service, though the company did not give a precise time frame for what it considered recent.

AI videos of dead celebrities are horrifying many of their families

bafkreicknb4ze3xnmdgiwdnqlf45kh2rukzgvbilpgbbhpq6mxr57b57iu

But another area where Sora users have been pushing the limits involves videos that reanimate dead celebrities.

And we’re not talking just JFK, MLK, and Einstein. Videos featuring more recently deceased figures such as Robin Williams (11 years ago), painter Bob Ross (30 years ago), Stephen Hawking (seven years ago), and even Queen Elizabeth II (three years ago) have been generated. Some of the videos are racist and offensive, shocking the relatives of the figures.

OpenAI told The Washington Post that it is now allowing representatives of “recently deceased” celebrities and public figures to request that their likenesses be blocked from the service, though the company did not give a precise time frame for what it considered recent.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.

©2025 Sherwood Media, LLC