Meta is “working with the public sector to adopt Llama across the US government,” according to CEO Mark Zuckerberg.
The comment, made during his opening remarks for Meta’s Q3 earnings call on Wednesday, raises a lot of important questions: Exactly which parts of the government will use Meta’s AI models? What will the AI be used for? Will there be any kind of military-specific applications of Llama? Is Meta getting paid for any of this?
When I asked Meta to elaborate, spokesperson Faith Eischen told me via email that “we’ve partnered with the US State Department to see how Llama could help address different challenges — from expanding access to safe water and reliable electricity, to helping support small businesses.” She also said the company has “been in touch with the Department of Education to learn how Llama could help make the financial aid process more user friendly for students and are in discussions with others about how Llama could be utilized to benefit the government.”
She added that there was “no payment involved” in these partnerships.
Advertisement
There’s also the cozying up to the government that Meta’s AI rivals are doing. OpenAI and Anthropic recently said they would share their models with the US AI Safety Institute ahead of time for safety screening. Google’s on-and–off-again relationship as an AI vendor for the Pentagon is well documented. In a recent blog post, OpenAI said its models were being used by DARPA, the U.S. Agency for International Development, and the Los Alamos National Laboratory.
While we wait to learn about Meta’s AI work with the government, Zuckerberg teased a bit more about the next Llama model on the Q3 earnings call. He said version four is training on “a cluster bigger than I’ve seen reported for anything else others are doing” and that he expects “new modalities,” “stronger reasoning,” and “much faster” performance when it debuts next year.
He acknowledged that Meta plans to continue spending more on AI in 2025, which is “maybe not want investors want to hear in the near term.” But he sees the upside as being worth it.
“I’m pretty amped about all the work we’re doing right now,” he said. “This may be the most dynamic moment I’ve seen in our industry, and I’m focused on making sure that we build some awesome things and make the most of the opportunities ahead.”
Advertisement
As a business, Meta is still continuing to grow. The company reported revenue of $40.5 billion for Q3, a 19-percent increase from a year ago, and $17.3 billion in profit. And it claims that 3.29 billion people use at least one of its apps each day, an increase of 5 percent from a year ago.
Update, October 30th: Added more details from Meta spokesperson.
It probably won’t come as much of a surprise to hear new research has found British workers often have to give up their spare time for work related commitments – but if you’re one of the 1 in 3 British workers who regularly skips their lunch break to catch up on admin tasks, HP has new tools for you.
The company has announced its new Amplify AI programme, which is set to offer “tools, resources, training, and certification” to make sure partners are well equipped to use AI to help their customers and within their own business.
An earlier survey, found that just 42% of workers and 51% of business leaders believe that AI can alleviate tedious tasks and allow employees to focus on the more critical aspects of their work – but HP hopes this will improve.
An AI PC for your admin needs
Alongside the Amplify programme, the company HP is releasing the new HP OmniBook Ultra 14 Flip, its next-gen AI PC 2-in-1 laptop costing £,1699 and featuring an inbuilt AI companion which can answer questions, analyze documents, and gather insights to help reduce workloads.
Advertisement
“AI is set to be a significant driver of change in the future of work, increasing productivity and allowing workers to enjoy their jobs more by taking simple and repetitive tasks off their plates,” said Neil Sawyer, Managing Director of HP’s Northwest Europe market.
“In the UK, according to HP’s WRI, 65% of workers believe AI will make their jobs easier, and at HP we want to make this a reality by empowering workers with the latest in AI technology to achieve their aspirations.” he continued.
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Advertisement
We have seen a slew of new products such as these, as Lenovo AI capable PCs saw a 228% rise in the second quarter of 2024. These products are hoping to encourage workers to adopt AI for mundane tasks to boost productivity across the board.
Currently, ChatGPT is the dominant service in the AI-powered chatbot segment. OpenAI, its parent company, was the one who started the AI-powered revolution of the tech industry after releasing the first version. Over the years, ChatGPT has received new features and much more powerful reasoning capabilities. Now, ChatGPT is getting a new search feature that seems quite useful.
If you are a frequent ChatGPT user, you might have a rather long conversation history with the chatbot. You might also want to quickly locate a specific conversation to retrieve some data. However, until now, the only way to do this was to scroll through the entire chat. This method was quite cumbersome and impractical, something that clashes with OpenAI’s philosophy of making the use of AI services as simple as possible.
ChatGPT is getting a search feature to find data in your chat history
Now, OpenAI is rolling out a search feature for ChatGPT’s chat history. This means that you can now easily find specific interactions you had with the platform weeks, months, or even years ago, using just a keyword or phrase. Just like any other built-in chat search system, you’ll only need to remember a word that you or ChatGPT used during the session. Of course, you need to have the chat history option enabled.
Availability is limited to Plus and Teams plans
Search systems for chat window-based services are pretty useful, and it’s quite surprising that ChatGPT lacked this feature for so long. Most messaging platforms or social networks usually include a similar option. Although you don’t actually chat with other people on ChatGPT, its “problem-solving” nature also deserved a search feature. Fortunately, the wait is over, and now finding information in your old chats with ChatGPT will be much easier.
Advertisement
Currently, the ChatGPT search system is available to subscribers of the Plus and Team plans. Subscribers to the Enterprise and Edu tiers will receive it starting next week. Lastly, free users of the service will gradually access the feature starting next month.
It’s no secret that X has become an even bigger cesspool of misleading information, unchecked claims and flat-out falsities since Elon Musk took over. Two new reports from The Center for Countering Digital Hate (CCDH) and The Washington Post reveal that the safeguards Musk removed and replaced aren’t controlling X’s problems with misinformation.
The CCDH published a report on its investigation into X’s Community Notes feature, a user-driven reporting system in which anonymous users write and rate correction for misleading posts. Researchers took a sample of 283 misleading election posts from the social media platform that received proposed Community Notes between January 1 and August 25. The report says that 209 of those misleading sample posts did not show the Community Notes correction to all X users. Even more alarming, the 209 misleading posts in question racked up 2.2 billion views.
The Washington Post followed the CCDH’s report with its own investigation into X’s Community Notes feature and found that X’s problems with misinformation go far beyond the election.
Former President Donald Trump made the bold claim during his only presidential debate with Vice President Kamala Harris that Haitians were eating people’s pets in Springfield, Ohio. Moderator and ABC news anchor David Muir corrected Trump’s statement as false because no such cases were reported to local police or government entities. The fact checking website Politifact rated Trump’s claim its lowest false rating of “Pants on Fire.” That didn’t stop this falsehood from spreading across X among conservative-leaning users.
Advertisement
The Post found that an account called End Wokeness with a following of 3.1 million X users started disseminating the former President’s claim about Haitian immigrants. The post remained unchecked for four days until one Community Notes user flagged the post as incorrect, citing five different articles to back up the correction. Unfortunately, the note failed to garner enough votes to label the post as false and it went uncorrected. As of Wednesday, the post is still on @EndWokeness’ account with a Community Note where it’s racked up 4.9 million views.
Musk’s account hasn’t helped the problem. The Post reports that he’s become “one of the X users most often targeted with proposed Community Notes” with one of 10 posts receiving a proposed correction note.
The publication cited a July post from @elonmusk containing a manipulated video of Harris spouting about President Joe Biden’s “senility” and how she became the nominee because she’s “the ultimate diversity hire.” You know where this is going. There’s no Community Notes or correction and the post is still on X even though thousands of replies from other X users are pointing out that it’s a fake. The post has a whopping 136.6 million views.
“Community Notes maintains a high bar to make notes effective and maintain trust across perspectives, and thousands of election and politics related notes have cleared that bar in 2024,” Keith Coleman, VP of product at X, said in a statement. “In the last month alone, hundreds of such notes have been shown on thousands of posts and have been seen tens of millions of times. It is because of their quality that notes are so effective.” Coleman, who oversees Community Notes, pointed to previous academic research into the feature. That research includes studies that found posts with a Community Note were 60 percent less likely to be shared, and that Community Notes result in an 80 percent uptick in post deletions.
Advertisement
The CCDH is one of Musk and X’s most vocal opponents. The British non-profit continually monitors Musk’s account for false posts that failed to earn a Community Note, particularly when it comes to the presidential election. CCDH CEO Imran Ahmed said in August that X “is failing woefully to contain the kind of algorithmically-boosted incitement that we all know can lead to real world violence. X took the CCDH to court over claims the non-profit created a “scare campaign” to bring down its advertising revenue. A US district court judge dismissed the lawsuit in March.
Update October 30, 2024, 9 PM ET: This story has been updated to add a statement and additional information from X VP of Product Keith Coleman.
Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More
Of all the AI video models out there, arguably the one to see most success among mainstream creators and viewers — those outside the pro and amateur filmmaking community — is Pika.
Following this, brands with accounts on the social networks Instagram and TikTok — especially brands in cosmetics, skincare, and wellness — began using the effects, especially the “squish,” to advertise their services.
Advertisement
It even sparked a whole trend of creators trying the “Squish It” effect — or Pikaffect, as the company calls its AI presents — on their own videos.
“We’re trying to put fun at the forefront of AI—making it accessible not just for creators, but for anyone, from kids to grandparents,” said Matan Cohen-Grumi, Pika’s Founding Creative Director, in a video call interview with VentureBeat earlier this week.
To use the new and prior Pikaffects, users of Pika follow the same simple steps: visit Pika.art, sign in with a Google Account, Discord Account, Facebook/Meta account or email address, and then navigate to the bottom menu bar to add a new image.
After tapping the Image button marked with a paperclip icon (highlighted above in a screenshot) the user can take a new image or add a previously uploaded one from their device or cloud photo library.
Then, tapping the Pikaffects button marked by a magic wand (encircled above in the annotated screenshot), the user can pull up all 13 preset Pikaeffects.
Finally, the user can generate a video based on the screenshot by tapping the star button (encircled above in the annotated screenshot).
“What I would suggest, is for everyone to go to our website and try it out,” advocated Cohen-Grumi. “It’s so, so accessible.”
The creative director asserted that Pika’s effects only take a few seconds to generate a new video from a still image.
Advertisement
However, in VentureBeat’s limited tests, the site appeared overloaded with traffic and stalled for a while with some images failing to generate videos so far on the company’s free tier, which offers 150 credits to the user each month — enough for 10 videos (1 video costs 15 credits on Pika’s scale). There are also Standard, Pro, and Unlimited tiers for $10, $35, and $95 per month (20% discount when paid annually) with gradually increasing numbers of credits.
Asked about the time outs we experienced, Cohen-Grumi noted that Pika’s newfound success with Pikaffects had come with load bearing challenges.
“We had a lot, a lot of traffic, more than created on the launch, but everything was resolved very quickly,” he told VentureBeat.
And seeking to dispel notions Pika was competing on novelty over realism, he also asserted that Pika 1.5 “can deliver extremely realistic results with natural movement.”
Advertisement
As for what’s next for Pika — more Pikaffects for every major holiday or season of the year? — Cohen-Grumi played coy.
“We’re always working on the next thing, ensuring everything we release is fun and accessible for everyone,” he said.
VB Daily
Stay in the know! Get the latest news in your inbox daily
The Shell logo is displayed outside a petrol station in Radstock in Somerset, England, on Feb. 17, 2024.
Matt Cardy | Getty Images News | Getty Images
British oil giant Shell on Thursday posted a small year-on-year drop to a stronger-than-expected third-quarter profit, partly owing to a sharp drop in crude prices and to lower refining margins.
The energy company reported adjusted earnings of $6 billion for the July-September period, beating analyst expectations of $5.3 billion, according to estimates compiled by LSEG.
Advertisement
Shell posted adjusted earnings of $6.3 billion in the second quarter and $6.2 billion in the third quarter of 2023.
Shell said it will buy back a further $3.5 billion of its shares over the next three months, while holding its dividend unchanged at 34 cents per share.
Net debt came in at $35.2 billion at the end of the third quarter, down from $40.5 billion when compared to the same period last year.
Shares of the London-listed firm have fallen around 3% year-to-date.
Advertisement
Ahead of the firm’s third-quarter earnings, Shell warned that refining profit margins had dropped by more than 28% on a quarterly basis, while trading results for its chemicals and oil products division were expected to be lower.
British rival BP on Tuesday posted its weakest quarterly earnings in nearly four years, weighed down by lower refining margins.
BP reported underlying replacement cost profit, used as a proxy for net profit, of $2.3 billion for the third quarter. That beat analyst expectations — but reflected a steep drop when compared to the same period a year earlier.
Oil prices tumbled over 17% in the third quarter amid concerns over the outlook for global oil demand.
You must be logged in to post a comment Login