The Chinese company Tencent introduced an open-source AI model, called HunyuanWorld-Voyager, that turns photos and images into explorable 3D sequences.
This video diffusion framework achieved the highest overall score of 77.62 on Stanford University’s WorldScore benchmark, surpassing competitors including WonderWorld (72.69) and CogVideoX-I2V (62.15).
HunyuanWorld-Voyager builds upon Tencent’s earlier HunyuanWorld 1.0 model, released in July. The new system generates both RGB video and depth information simultaneously, allowing users to navigate virtual environments through keyboard or joystick controls.
According to Tencent’s announcement on September 2, HunyuanWorld-Voyager is the industry’s first ultra-long-range world model with native 3D reconstruction capabilities.
OpenAI announced on Thursday that it reached a tentative deal to settle financial issues with its partner Microsoft, which invested over $13 billion in the company in exchange for 49% of OpenAI’s future profits.
Microsoft and OpenAI provided no additional details of the deal, saying in a joint statement that they had signed a nonbinding memorandum of understanding to restructure their relationship and that they would look to formalize it in a contract.
Also on Thursday, OpenAI announced that it was providing an equity stake that will exceed 20 percent of the reorganized company and will be worth at least $100 billion to the nonprofit that controls the organization.
Both moves were considered essential steps in OpenAI’s shift from being managed by a nonprofit to becoming a public benefit corporation.
As its business has needed to raise more money, OpenAI has been trying to transition to a more traditional for-profit structure, specifically a Public Benefit Corporation (PBC). That might enable the company to eventually issue shares on the stock market and raise capital from the general public.
The nonprofit that now manages OpenAI would become “one of the most well-resourced philanthropic organizations in the world,” said Bret Taylor, OpenAI’s board chair in a blog post.
As part of this next phase, the OpenAI nonprofit has launched a call for applications for the first wave of a $50 million grant initiative to support nonprofit and community organizations in three areas: AI literacy and public understanding, community innovation, and economic opportunity.
OpenAI unveiled that it is planning to launch, by mid-2026, a LinkedIn-style hiring platform to connect businesses and employees.
This OpenAI Jobs Platform will “use AI to find the perfect matches between what companies need and what workers can offer,” OpenAI CEO of Applications Fidji Simo announced the new endeavor in a blog post on Thursday.
The San Francisco-based research lab seeks to expand beyond its core consumer offering. OpenAI is reportedly working on a browser and a social media app, too.
OpenAI’s job platform can put the company in direct competition with LinkedIn, which was co-founded by Reid Hoffman, one of OpenAI’s earliest investors, and is owned by Microsoft, OpenAI’s largest financial backer.
OpenAI also said it will start offering certifications for people with different levels of “AI fluency” by the end of this year through its OpenAI Academy and this initiative called OpenAI Certifications.
“OpenAI is committing to certifying 10 million Americans by 2030. And we’ll be doing it with our launch partners, including the biggest private employer in the world: Walmart,” said Fidji Simo.
The company said that these two initiatives are part of its commitment to the White House’s efforts toward expanding AI literacy.
AI-driven private schools, run by for-profit companies, are expanding their footprint in the U.S. while public schools struggle to attract and retain students, and others accelerate their investments in AI.
These schools teach core subjects for two hours a day and devote the afternoon to developing practical skills, such as financial literacy, public speaking, and entrepreneurship.
Alpha Schools was founded in 2014 as a pricey private school in Austin. Today, with an annual tuition of $45,000, it is leading the parental school choice movement while embracing AI technology that generates personalized learning plans for students.
This fall, Alpha Schools will launch in Santa Barbara, California; New York City; Chantilly, Virginia; and Raleigh and Charlotte, North Carolina, before expanding to Houston, Tampa, and Puerto Rico.
Existing locations are in Scottsdale, San Francisco, Miami, and Palm Beach. There are also five in Texas, including Brownsville, home of Elon Musk’s SpaceX.
Instead of teachers, these schools employ professional guides and coaches who
come from a range of backgrounds, from tech to law. Their mission is to motivate students. They make over $100,000 a year.
Students learn on third-party apps, such as Synthesis Tutor and Math Academy, as well as Alpha Schools’ own programs. Each subject is taught in 25-minute sessions, with short breaks in between.
Skeptics have questioned the effectiveness of replacing teachers with AI-assisted learning and restricting learning to 25 minutes per subject.
Alpha School is gaining more national attention, boosted by the support of billionaire Bill Ackman, an outspoken critic of DEI.
Through a splashy event yesterday, Apple introduced its newest iPhones: the iPhone 17, 17 Pro, 17 Pro Max, and a sleek, lightweight, and slimmer version, the iPhone Air, [in the picture above], 5.6 mm with a 6.5-inch display device.
The Air’s price point of $999 is 22% more expensive than the 17 base model, which starts at $799. The iPhone 17 Pro starts at $1,099.
At Apple’s event, the audience witnessed that the company continues to market itself as a best-in-class hardware maker first, not an AI device maker.
The most compelling use of AI was the Live Translation feature coming to AirPods 3. Priced at $249, AirPods 3 feature live translation technology, powered by Apple Intelligence, to help users translate foreign languages in real time.
Beyond the AirPods, AI technology has received only minor upgrades, mostly to its front camera. There was no mention of Siri.
To date, Apple has only released a baseline of AI features, such as AI writing tools, summarization, generative AI images, live translation, visual search, and Genmoji.
Sources said that Apple is still looking to outsource some technology to Google Gemini to catch up in the AI race.
Meanwhile, Google last month rolled out its latest release of an AI-powered Android phone with its Pixel 10.
TechCrunch wrote that “today’s iPhone owners often swap out Apple’s technology for Google’s by opting for Gmail, Google Drive and Docs, Google Maps, and Chrome over Apple’s own apps like Mail, its iWork suite, Apple Maps, and Safari, for example. When people search the web, they turn to Google’s Search app, not Apple’s built-in Spotlight search, despite its many integrations over the years to offer basic facts and answers, leveraging sources like Wikipedia.”
The company conveyed that the look and feel of updated iPhones, along with their hardware advancements, will continue to drive sales, enabling Apple to incorporate quality, camera improvements, privacy-preserving technology, intentional software design changes like Liquid Retina, and now, super-thin phones.
first new iphone upgrade i have really wanted in awhile! looks very cool.
Google launched three Veo updates: support for vertical format outputs (9:16 aspect ratio), 1080p HD output, and a price reduction.
The company is also making Veo 3 and Veo 3 Fast stable and ready for scaled production use in the Gemini API.
The lower price on Veo 3 is $0.40 per second, compared to $0.75 per second previously. In Veo 3, Fast is now $0.15 per second, when it was $0.40 per second.
The vertical video and 1080p formats are now supported, making it suitable for mobile-first and social use cases, and allowing users to enjoy higher quality content, respectively.
French start-up Mistral AI introduced a memory system on its Le Chat assistant, enabling the model to remember, adapt, and stay on task.
The company also unveiled a connector directory built on MCP, supporting more than 20 enterprise platforms, spanning data platforms (Databricks, Snowflake, Pinecone), productivity tools (Notion, Box, Asana), development platforms (GitHub, Linear, Sentry), and financial services (Stripe, PayPal, Square, Plaid).
Both capabilities will be available to free users, an offer that undercuts competitors such as ChatGPT and Claude, who reserve such features for premium subscribers.
In addition, unlike some competitors, Mistral emphasized user control over stored information in Le Chat, allowing users to add, edit, or delete memory entries at any time, thereby addressing privacy concerns. The company also supports importing memories from ChatGPT, potentially easing migration for users considering a switch. Memories, therefore, are portable and interoperable by design, with users owning their memories and being able to export or import them from elsewhere.
Currently, many AI models store information automatically, but only recall it when the users explicitly ask, or worse, resurface it without warning. Memory keeps chatbots anchored in context, retaining user preferences and past decisions, delivering more personalized and relevant responses over time.
Mistral AI said its memory system is “designed to help users think better, not guess what they are doing.” The company explained, “Le Chat clearly shows when it’s recalling something, where it came from, and why it’s relevant.”
Memory Insights’ lightweight prompts help users explore what Le Chat remembers and how it can assist them. They surface trends, suggest summaries, and point out moments worth revisiting, all based on users’ data, and all editable. It’s a way to turn memory from passive storage into an active signal.
The company stated that it would continue to enhance memory capabilities: trimming noise, speeding up recall, and making it easier to organize long-term information.
President Trump hosted a dinner for the leaders of the world’s biggest technology companies at the White House on Thursday night, receiving their unanimous praise.
Each participant — including Apple Chief Executive Tim Cook and OpenAI CEO Sam Altman — took a turn thanking Donald J. Trump for his efforts to promote investments in chip manufacturing and AI while laying out how much their companies plan to invest in the U.S.
“Thank you for being such a pro-business, pro-innovation president. It’s a very refreshing change,” Altman said. “I think it’s going to set us up for a long period of leading the world, and that wouldn’t be happening without your leadership.”
Tim Cook said Apple is expected to invest $600 billion in the U.S. “I want to thank you for setting the tone such that we can make a major investment in the United States and have some key manufacturing here. I think it says a lot about your leadership and focus on innovation,” Cook said.
Oracle CEO Safra Catz was the most enthusiastic. “You’ve unleashed American innovation and creativity. All the work you’re doing in basically every cabinet post, in addition to what’s coming out of the White House, is making it possible for America to win,” Catz said. “I think this is the most exciting time in America ever.”
Tech titans present included Alphabet and Google CEO Sundar Pichai and IBM CEO and Chairman Arvind Krishna. Meta Platforms CEO Mark Zuckerberg was seated directly to the president’s right. Notably from the gathering was Elon Musk.
“The most brilliant people are gathered at this table,” Trump said. “This is definitely a high-IQ group.”
The model was developed by specialized engineers and a large number of researchers of ETH Zurich, the Swiss Federal Institute of Technology Lausanne (EPFL), and the Swiss National Supercomputing Centre (CSCS).
The model’s name, ‘Apertus,’ derived from the Latin word for “open,” highlights its open-source architecture, training data, model weights, and intermediate checkpoints. It comes in two sizes: 8 billion, for individual use, and 70 billion parameters.
The model, unveiled this week, is now accessible and documented on Hugging Face, serving as a building block for developers and organizations to create applications such as chatbots, translation systems, or educational tools.
As of today, only Swisscom business customers are able to access the Apertus model via its AI platform.
“We see it as a driver of innovation and a means of strengthening AI expertise across research, society, and industry,” said Thomas Schulthess, Director of CSCS and Professor at ETH Zurich.
“Apertus is built for the public good. It stands among the few fully open LLMs at this scale and is the first of its kind to embody multilingualism, transparency, and compliance as foundational design principles”, added Imanol Schlag, technical lead of the LLM project and Research Scientist at ETH Zurich.
The authors of this LLM said, without providing further detail, that “upcoming Swiss {ai} Weeks hackathons will be the first opportunity for developers to experiment hands-on with Apertus, test its capabilities, and provide feedback for improvements to future versions.”
Meta rolled out this month an AI feature that allows users on Instagram and Facebook to translate voice content with their sound, tone, and lip-sync from English to Spanish.
With more languages to be added over time, these AI translations are available to Facebook creators with 1,000 or more followers and all public Instagram accounts globally.
Creators can view translations and lip syncs before they’re posted publicly. Viewers watching the translated reel will see a notice at the bottom that indicates it was translated with Meta AI.
Creators are also gaining access to a new metric in their Insights panel, where they can see their views by language.
Instagram head Adam Mosseri explained in a post on Instagram, “If we can help you reach those audiences who speak other languages, reach across cultural and linguistic barriers, we can help you grow your following and get more value out of Instagram and the platform.”
![Apple Marketed Its New iPhones As a Best-In-Class Hardware, Not As an AI Device Maker [Video]](https://cms.iblnews.org/wp-content/uploads/2025/09/iPhoneair.jpg)
![Tech CEOs Praised President Trump at White House Dinner [Video]](https://cms.iblnews.org/wp-content/uploads/2025/09/techleaderswithtrump.jpg)
