Intel [1]shelled out $16.5 billion on R&D in 2024 , outspending Nvidia by 28% and AMD by 156%, with much of the cash going into chip design, fabrication tech, and the upcoming Nova Lake architecture. "When you compare the R&D expenditures to the amount of revenue, though, the story takes on a very different look," notes PC Gamer. "Intel spent 31% of its net revenue, and 26% for AMD, but Nvidia and Samsung got by on just 10% and 4%, respectively." From the report:

> An analysis of research and development expenditure by TechInsights was reported by [2]Korea JoongAng Daily , but you can get the numbers yourself by pulling up each company's 2024 financial results. For example, AMD declared that it spent $6.456 billion last year ( [3]pdf , page 1) on R&D, whereas Nvidia [4]forked out $12.914 billion. It's worth noting that Nvidia's financial statements are numbered one year ahead of the actual period (FY 2026 is 2025 and so on).

>

> Anyway, those figures pale in comparison to how much cash Intel burned through in 2024 to research and develop chip, fabrication technologies, software, and all kinds of tech stuffâ"a staggering $16.546 billion ( [5]pdf , page 25). That's 28% more than Nvidia and a frankly unbelievable 156% more than AMD. The nearest non-US semiconductor firm is Samsung Electronics, which spent a reported $9.5 billion on R&D. That would place third, comfortably ahead of AMD, and it strongly suggests that if you have your own foundries for making chips, you need to spend a lot of cash on finding ways to make better processors.



[1] https://www.pcgamer.com/hardware/intel-spent-so-much-cash-on-research-and-development-last-year-that-it-outspent-nvidia-by-28-percent-and-amd-by-a-whopping-156-percent/

[2] https://koreajoongangdaily.joins.com/news/2025-09-02/business/tech/Samsung-hikes-RD-spending-a-worldtopping-71-percent/2389595

[3] https://d1io3yog0oux5.cloudfront.net/_0fe5f5a4cf1d169c12435555bc42ec4f/amd/db/841/9090/financial_tables_pdf/Q4+2024+Gaap+and+Non-Gaap+Earnings+Tables.pdf

[4] https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-fourth-quarter-and-fiscal-2025

[5] https://www.intc.com/filings-reports/all-sec-filings/content/0000050863-25-000009/0000050863-25-000009.pdf

Philips Hue is rolling out MotionAware, a new feature that [1]turns its smart bulbs into motion sensors using radio-frequency (RF) Zigbee signals . The upgrade works with most Hue bulbs made since 2014, but requires the new $99 Bridge Pro hub to enable. The Verge reports:

> To create a MotionAware motion-sensing zone, you need Hue's new Bridge Pro and at least three Hue devices in a room. It works with all new and most existing mains-powered Hue products via a firmware update. That includes smart bulbs, light strips, and fixtures. Portable devices, such as the Hue Go or Table Lamp, and battery-powered accessories, such as Hue switches, aren't compatible. Neither is Hue's current smart plug. [...] "All of the functionality you get with our physical motion sensors -- including turning on when motion is detected or off when there's been no movement for a certain amount of time -- can be configured on motion-aware motion events," says George Yianni, Hue CTO and founder, in an interview with The Verge. "We've done something that's quite a lot better than what else is out there."

>

> MotionAware is occupancy sensing, not presence sensing; it requires movement. Yianni says it's comparable to the passive infrared sensing (PIR) Hue's physical sensors use. This means it can be triggered by pets or other motion. A sensitivity slider in the app helps fine-tune detection. According to Yianni, a key benefit over PIR is that a MotionAware zone can cover a larger area than a single PIR sensor, and it's also not limited to line of sight. MotionAware can't sense light levels, which Hue Motion Sensors can, but you can pair a light sensor to a motion zone to feed it that data. The positioning of the lights will also play a role in determining the effectiveness of the motion sensing. "We recommend that the lights surround an area which will roughly define the detection area in which motion will be detected," says Yianni. "It will sense around the lights and in the broader room thanks to reflections, but detection reliability will depend on lots of factors."

>

> Beyond lighting automation, MotionAware can also integrate with Hue Secure, Hue's DIY security platform that includes cameras, contact sensors, and a new video doorbell. Motion detection can trigger lights to flash red, activate Hue's new plug-in chime/siren, and send an alert to your phone with a button to call emergency services. [...] MotionAware is built on RF sensing -- a technology that uses wireless signals to "see" a space and detect disruptions within it. The data is then sent to the Bridge Pro, where AI algorithms are applied to figure out what is causing those disruptions, so the system can act accordingly. This is why it's limited to the Bridge Pro, the V2 bridge isn't powerful enough to run those algorithms, says Yianni.



[1] https://www.theverge.com/smart-home/768585/philips-hue-motion-aware-rf-sensing-lights-motion-ivani

An anonymous reader quotes a report from The Guardian:

> Managers and supervisors brace yourselves: calling the boss a dickhead is [1]not necessarily a sackable offense , a tribunal has ruled. The ruling came in the case of an office manager who was sacked on the spot when -- during a row -- she called her manager and another director dickheads. Kerrie Herbert has been awarded almost 30,000 pounds in compensation and legal costs after an employment tribunal found she had been unfairly dismissed.

>

> The employment judge Sonia Boyes ruled that the scaffolding and brickwork company she worked for had not "acted reasonably in all the circumstances in treating [her] conduct as a sufficient reason to dismiss her." "She made a one-off comment to her line manager about him and a director of the business," Boyes said. "The comment was made during a heated meeting. "Whilst her comment was not acceptable, there is no suggestion that she had made such comments previously. Further ... this one-off comment did not amount to gross misconduct or misconduct so serious to justify summary dismissal." [...]

>

> Boyes found that Herbert was summarily fired because of her use of the word "dickheads" and ruled that the company had failed to follow proper disciplinary procedures. She concluded that calling her bosses dickheads was not sufficient to fire Herbert and ordered the firm to pay 15,042.81 pounds in compensation. In her latest judgment she also ruled it had to pay 14,087 pounds towards her legal fees.

"If it was anyone else in this position they would have walked years ago due to the goings-on in the office, but it is only because of you two dickheads that I stayed," said Herbert.

Swannell retorted: "Don't call me a fucking dickhead or my wife. That's it, you're sacked. Pack your kit and fuck off."



[1] https://www.theguardian.com/money/2025/sep/04/calling-your-boss-a-dickhead-is-not-a-sackable-offence-tribunal-rules

Nepal's government said Thursday it is [1]blocking most social media platforms including Facebook, X and YouTube because the companies failed to comply with regulations that required them to register with the government. From a report:

> Nepal's Minister for Communication and Information Prithvi Subba Gurung said about two dozen social network platforms that are widely used in Nepal were repeatedly given notices to come forward and register their companies officially in the country. The platforms would be blocked immediately, he said.

>

> TikTok, Viber and three other social media platforms would be allowed to operate in Nepal because they have registered with the government. Nepal government have been asking the companies to appoint a liaison office or point in the country. It has brought a bill in parliament that aims to ensure that social platforms are properly managed, responsible and accountable.



[1] https://apnews.com/article/nepal-ban-social-media-platform-3b42bbbd07bc9b97acb4df09d42029d5

Germany has already [1]met its 2028 goal for reducing coal-fired power generation , so won't need to order the shutdown of any plants for a second year running, the country's regulator said. From a report:

> Germany has an interim 2028 target of reducing coal-fired power by 8.7 gigawatts, and as of Sept. 1 it had exceeded this level by about 10%, the Federal Network Agency said on its website on Monday.

>

> Almost two thirds of Germany's electricity comes from renewables and excess solar power production has frequently pushed prices below zero, making burning coal less profitable. Yet Europe's largest economy remains heavily dependent on the fossil fuel and is still the European Union's biggest polluter.



[1] https://www.bloomberg.com/news/articles/2025-09-01/germany-already-met-its-2028-goal-for-reducing-coal-fired-power

OpenAI plans to launch [1]a new AI-powered jobs platform next year to [2]help match employers with candidates who have AI skills in a bid to accelerate the technology's deployment across businesses and government agencies. From a report:

> The ChatGPT maker will also introduce a new certification program in the coming months that will teach workers how to better use AI on the job. OpenAI is working with multiple organizations on the program, including Walmart, the largest private employer in the US. OpenAI said it plans to certify 10 million Americans by 2030. [...] For the jobs platform, OpenAI plans to use AI to help match local governments and companies of all sizes with potential candidates.



[1] https://openai.com/index/expanding-economic-opportunity-with-ai/

[2] https://www.bloomberg.com/news/articles/2025-09-04/openai-unveils-jobs-platform-certification-program-for-ai-roles

An anonymous reader shares a report:

> Adobe is bringing its video editor Premiere to iPhone, [1]promising "pro-level" editing on the go for free . The app will launch later this month, with an Android version also under development.

>

> The Premiere app features a familiar multi-track timeline, with support for an unlimited number of video, audio, and text layers. There's automatic captioning, 4K HDR support, and one-tap exporting to TikTok, YouTube Shorts, and Instagram -- including automatic resizing that frames content for each platform.

>

> The iOS version of Premiere will be free to download and use, though Adobe says there will be charges for additional cloud storage and generative AI credits. Speaking of, it includes support for Adobe's generative sound effects and AI-powered speech enhancement, plus a wider range of AI assets generated through Adobe Firefly. If you'd rather not use AI content, there's a selection of Adobe fonts, along with images, sounds, music, and video assets that are free to use.



[1] https://www.theverge.com/news/771236/adobe-premiere-iphone-ios-video-editor-android

Tom's Hardware:

> According to an early Amazon Mexico listing, Logitech is preparing to launch the Signature Slim Solar+, a wireless keyboard in every sense of the word. The Signature Slim Solar+ [1]features a solar panel , providing owners with a battery lasting up to 10 years.



[1] https://www.tomshardware.com/peripherals/keyboards/solar-powered-logitech-keyboard-appears-on-amazon-mexico-mx-keys-s-look-alike-promises-up-to-10-years-of-power

Nature's editorial board argues the UN's upcoming trans fat elimination proposal must specify "industrially produced" fats [1]to avoid unintended consequences for nutrition in poor countries . The board supports a coalition led by the International Livestock Research Institute and African Union requesting precise language, noting natural trans fats occur in milk and dairy products.

The editorial extends criticism to the broader ultra-processed foods classification system. While acknowledging the 2009 NOVA scale by University of Sao Paulo's Carlos Monteiro pushed governments toward strong public health policies, Nature says the ultra-processed category problematically groups baby formula with hot dogs. The board calls for improved definitions that balance reining in industrial food production excesses while ensuring adequate calorie access globally.



[1] https://www.nature.com/articles/d41586-025-02750-0

An anonymous reader [1]shares a report :

> The total number of GPUs sold for the second quarter of 2025 hit 11.6 million units, while desktop PC CPUs went up to 21.7 million units, according to a Jon Peddie Research report. This is a 27% increase in graphics card shipments and a 21.6% jump in CPU shipments from the last quarter, which is a change from the usual drop in deliveries we've seen in recent years.

>

> "AIB prices dropped for midrange and entry-level, while high-end AIB prices increased, and most retail suppliers ran out of stock. This is very unusual for the second quarter," said Jon Peddie Research president Dr. Jon Peddie. "We think it is a continuation of higher prices expected due to the tariffs and buyers trying to get ahead of that."

>

> As for the three major GPU manufacturers, Nvidia still has the lead, taking in 94% of the market -- an increase of 2.1% over the previous quarter -- while AMD is at a distant second place with 6%. This is still a much better position than Intel, though, whose market share is so small it did not even register on the chart.



[1] https://www.tomshardware.com/pc-components/gpus/nvidia-dominates-gpu-shipments-with-94-percent-share-shipment-surge-likely-caused-by-customers-getting-ahead-of-tariffs

[1]alternative_right writes:

> For decades, fragments and unofficial copies of Microsoft's 6502 BASIC have circulated online, mirrored on retrocomputing sites, and preserved in museum archives. Coders have studied the code, rebuilt it, and even run it in modern systems. Today, for the first time, we're opening the hatch and officially [2]releasing the code under an open-source license . Microsoft BASIC began in 1975 as the company's very first product: a BASIC interpreter for the Intel 8080, written by Bill Gates and Paul Allen for the Altair 8800. That codebase was soon adapted to run on other 8-bit CPUs, including the MOS 6502, Motorola 6800, and 6809.

>

> The 6502 port was completed in 1976 by Bill Gates and Ric Weiland. In 1977, Commodore licensed it for a flat fee of $25,000, a deal that placed Microsoft BASIC at the heart of Commodore's PET computers and, later, the VIC-20 and Commodore 64. The version we are releasing here -- labeled "1.1" -- contains fixes to the garbage collector identified by Commodore and jointly implemented in 1978 by Commodore engineer John Feagans and Bill Gates, when Feagans traveled to Microsoft's Bellevue offices. This is the version that shipped as the PET's "BASIC V2." It even contains a playful Bill Gates Easter egg, hidden in the labels STORDO and STORD0, which Gates himself confirmed in 2010.



[1] https://slashdot.org/~alternative_right

[2] https://opensource.microsoft.com/blog/2025/09/03/microsoft-open-source-historic-6502-basic/

Rising adoption of AI technology by firms in the Federal Reserve's New York district [1]has not been much of a job-killer so far , the regional Fed bank said in a blog on Thursday. Reuters:

> "Businesses reported a notable increase in AI use over the past year, yet very few firms reported AI-induced layoffs," New York Fed economists wrote in the blog. "Indeed, for those already employed, our results indicate AI is more likely to result in retraining than job loss, similar to our findings from last year," and so far the technology does not point to "significant reductions in employment."

>

> There has been broad concern that AI could create major headwinds for hiring in the coming years, with the technology hitting highly-paid professional and managerial jobs the hardest. Investors are plowing cash into AI investments at a time when employment has already begun to show some softness, although job market changes related to AI will almost certainly play out over a long time horizon. The New York Fed blog noted that the modest impact on jobs so far may not hold in the future. "Looking ahead, firms anticipate more significant layoffs and scaled-back hiring as they continue to integrate AI into their operations," New York Fed researchers wrote.



[1] https://money.usnews.com/investing/news/articles/2025-09-04/ai-not-affecting-job-market-much-so-far-new-york-fed-says

AI will disrupt the labor market within five to ten years and force Americans to [1]rely on investment returns rather than wages for income , according to Vlad Tenev, chief executive of stock trading firm Robinhood. Tenev told Fortune that "if you can't rely on labor to generate money to make a living, capital becomes more important."

The brokerage chief said private companies and government must make investing easier from an early age. He cited the proposed Invest America Act, included in congressional reconciliation legislation, which would provide every newborn with $1,000 in an investment account. Tenev said the policy represents preparation for an economy where "humans comprise less than 1% of the total intelligence" as AI systems advance beyond current capabilities.



[1] https://fortune.com/2025/08/27/robinhood-ceo-vlad-tenev-leadership-next/

Atlassian said it has [1]agreed to acquire The Browser Co. , a startup that offers a web browser with AI features, for $610 million in cash. CNBC:

> The companies aim to close the deal in Atlassian's fiscal second quarter, which ends in December. Established in 2019, The Browser Co. has gone up against some of the world's largest companies, including Google, with Chrome, and Apple, which includes Safari on its computers running MacOS. The startup debuted Arc, a customizable browser with a built-in whiteboard and the ability to share groups of tabs, in 2022.

>

> The Dia browser, a simpler option that allows people to chat with an AI assistant about multiple browser tabs at once, became available in beta in June. Atlassian co-founder and CEO Mike Cannon-Brookes said he sees shortcomings in the most popular browsers for those who do much of their work on computers.

Further reading : [2]Atlassian Buying The Browser Company Feels Like a Waste of Money .



[1] https://www.cnbc.com/2025/09/04/atlassian-the-browser-company-deal.html

[2] https://nerds.xyz/2025/09/atlassian-browser-company-acquisition-opinion/

Investment research firm Bernstein warned Thursday that India [1]faces a "strategic tech crisis" as US technology giants deploy predatory pricing strategies to lock up the Indian AI market. Perplexity Pro launched [2]free for one year to Airtel's 350 million subscribers while OpenAI [3]introduced a $5 monthly India subscription compared to $20 in the United States.

Bernstein analysts described regulatory "double standards" where foreign tech companies receive favorable treatment while domestic companies face what the firm called "crushing rules and government-led 'tech stacks' that make private business unviable." Private AI investment in India totaled $11.29 billion between 2013 and 2024 compared to $471 billion in the United States and $119 billion in China. From the report:

> When OpenAI, which is reportedly looking to set up a data center in India, announced the plans to launch a new office, it was met with another round of excitement -- "as if Open AI will hire all Indians at hefty salaries," the firm wrote in a note to clients Thursday. Bernstein analysts pour cold water on this excitement, dismissing it as a "repeat of the 90s" and arguing that the hype misses the fundamental power imbalance.

>

> "Anyone, we repeat anyone, can build a data center... This is the start of the dominance of US tech in Indian AI environment ensuring Indian entrepreneurs do not get a fighting chance to stay relevant. They will run on the sidelines - piggybacking on the US foundation models or maybe even the Chinese," they wrote.



[1] https://indiadispatch.com/p/bernstein-is-not-buying-india-ai-story

[2] https://indiadispatch.com/p/perplexity-pro-airtel-free-india

[3] https://slashdot.org/story/25/08/19/0333218/openai-launches-46-budget-ai-subscription-tier-in-india

An anonymous reader quotes a report from Ars Technica:

> On Tuesday, Tencent [1]released HunyuanWorld-Voyager, a new open-weights AI model that [2]generates 3D-consistent video sequences from a single image , allowing users to pilot a camera path to "explore" virtual scenes. The model simultaneously generates RGB video and depth information to enable direct 3D reconstruction without the need for traditional modeling techniques. However, it won't be replacing video games anytime soon.

>

> The results aren't true 3D models, but they achieve a similar effect: The AI tool generates 2D video frames that maintain spatial consistency as if a camera were moving through a real 3D space. Each generation produces just 49 frames -- roughly two seconds of video -- though multiple clips can be chained together for sequences lasting "several minutes," according to Tencent. Objects stay in the same relative positions when the camera moves around them, and the perspective changes correctly as you would expect in a real 3D environment. While the output is video with depth maps rather than true 3D models, this information can be converted into 3D point clouds for reconstruction purposes.

There are some caveats with the tool. It doesn't generate true 3D models (only 2D frames with depth maps) and each run produces just two seconds of footage, with errors compounding during longer or complex camera motions like full 360-degree rotations. Furthermore, because it relies heavily on training data patterns, its ability to generalize is limited and it demands enormous GPU power (60-80GB of memory) to run effectively. On top of that, [3]licensing restricts use in the EU, UK, and South Korea, with large-scale deployments requiring special agreements.

Tencent published the model weights on [4]Hugging Face .



[1] https://github.com/Tencent-Hunyuan/HunyuanWorld-Voyager

[2] https://arstechnica.com/ai/2025/09/new-ai-model-turns-photos-into-explorable-3d-worlds-with-caveats/

[3] https://github.com/Tencent-Hunyuan/HunyuanWorld-Voyager/blob/main/LICENSE

[4] https://huggingface.co/tencent/HunyuanWorld-Voyager

A long-term study of over 12,000 adults suggests that artificial sweeteners like aspartame, saccharin, and sugar alcohols [1]may accelerate cognitive decline in middle age , equivalent to about 1.6 years of extra aging. The Guardian reports:

> Sweeteners' association with cognitive decline is of such concern that consumers should instead use either tagatose, a natural sweetener, or alternatives such as honey or maple syrup, the researchers said. They looked at the impact of seven sweeteners on the health of the study's participants -- 12,772 civil servants in Brazil, with an average age of 52 -- who were followed up for on average eight years. Participants completed questionnaires detailing their food and drink intake over the previous year, and later underwent tests of their cognitive skills such as verbal fluency and word recall.

>

> People who consumed the most sweeteners experienced declines in their thinking and memory skills 62% faster than those with the lowest intake, the researchers found. This was "the equivalent of about 1.6 years of aging," the researchers said. Consumption of combined and individual LNCs, particularly aspartame, saccharin, acesulfame K, erythritol, sorbitol and xylitol, was associated with cognitive loss. "Daily consumption of LNCs was associated with accelerated decline in memory, verbal fluency and global cognition," the authors say in their paper, [2]published in the American medical journal Neurology . However, the trend was only observed in participants under the age of 60. That shows that middle-aged adults need to be encouraged to use fewer sweeteners, they added.



[1] https://www.theguardian.com/food/2025/sep/03/sweeteners-can-harm-cognitive-health-equivalent-to-16-years-of-ageing-study-finds

[2] https://www.neurology.org/doi/10.1212/WNL.0000000000214023

After 26 years and over 1,000 episodes, Melvyn Bragg is [1]stepping down as presenter of BBC Radio 4's In Our Time , leaving behind a legacy of intellectual curiosity and broadcasting excellence. While he will no longer host the series, he will remain involved with the BBC and is set to launch a new project in 2026. The BBC reports:

> Over the last quarter of a century, Melvyn has skilfully led conversations about everything from the age of the Universe to 'Zenobia', Queen of the Palmyrene Empire. He has welcomed the company of the brightest and best academics in their fields, sharing their passion and knowledge with a fascinated audience right around the globe. While he will be much missed on [2]In Our Time , Melvyn will continue to be a friend of Radio 4 with more to come to celebrate his extraordinary career, and a new series in 2026 (details to be announced soon).

>

> Melvyn Bragg says: "For a program with a wholly misleading title which started from scratch with a six-month contract, it's been quite a ride! I have worked with many extremely talented and helpful people inside the BBC as well as some of the greatest academics around the world. It's been a great privilege and pleasure. I much look forward to continuing to work for the BBC on Radio 4. Thank you for listening." [...] In Our Time will be back on Radio 4 with a new presenter who will be announced in due course.



[1] https://www.bbc.co.uk/mediacentre/2025/melvyn-bragg-decides-to-step-down-from-presenting-in-our-time

[2] https://www.bbc.co.uk/programmes/b006qykl/episodes/player

An anonymous reader quotes a report from The Hill:

> Russian President Vladimir Putin and Chinese President Xi Jinping were caught on a hot mic [1]discussing organ transplants and immortality at the military parade in Beijing on Wednesday. The two leaders were [2]captured on the stream as they walked with North Korean leader Kim Jong Un at Tiananmen Square, with the Russian translator saying: "Biotechnology is continuously developing," according to [3]Reuters . "Human organs can be continuously transplanted. The longer you live, the younger you become, and (you can) even achieve immortality," the translator added.

>

> Xi responded by saying that some predict that humans could live up to "150 years old." The Kremlin head later confirmed that the two leaders discussed immortality. "Modern means of healing, and medical means, all kinds of surgical means related to organ replacement, they allow humanity to hope that active life will continue not as it does today. The average age in different countries is different, but nevertheless, life expectancy will increase significantly," Putin told reporters, according to [4]CNN .



[1] https://thehill.com/policy/international/5484700-putin-xi-caught-on-hot-mic-discussing-organ-transplants-and-immortality/

[2] https://www.cnn.com/2025/09/03/world/video/putin-xi-hot-mic-immortality-vrtc

[3] https://www.reuters.com/business/media-telecom/hot-mic-picks-up-putin-xi-discussing-organ-transplants-immortality-2025-09-03/

[4] https://www.cnn.com/2025/09/03/asia/xi-putin-chat-immortality-longevity-intl

Switzerland has [1]launched Apertus, a fully open-source, multilingual LLM trained on 15 trillion tokens and over 1,000 languages. "What distinguishes Apertus from many other generative AI systems is [2]its commitment to complete openness ," reports CyberInsider. From the report:

> Unlike popular proprietary models, where users can only interact via APIs or hosted interfaces, Apertus provides open access to its model weights, training datasets, documentation, and even intermediate checkpoints. The source code and all training materials are released under a permissive open-source license that allows commercial use. Since the full training process is documented and reproducible, researchers and watchdogs can audit the data sources, verify compliance with data protection laws, and inspect how the model was trained. Apertus' development explicitly adhered to Swiss data protection and copyright laws, and incorporated retroactive opt-out mechanisms to respect data source preferences.

>

> From a privacy perspective, Apertus represents a compelling shift in the AI landscape. The model only uses publicly available data, filtered to exclude personal information and to honor opt-out signals from content sources. This not only aligns with emerging regulatory frameworks like the EU AI Act, but also provides a tangible example of how AI can be both powerful and privacy-respecting. [3]According to ETH Zurich's Imanol Schlag , technical lead of the project at ETH Zurich, Apertus is "built for the public good" and is a demonstration of how AI can be deployed as a public digital infrastructure, much like utilities or transportation.

The model is available via Swisscom's [4]Sovereign Swiss AI Platform . It's also available through [5]Hugging Face and the [6]Public AI Inference Utility .



[1] https://www.swiss-ai.org/apertus

[2] https://cyberinsider.com/switzerland-launches-apertus-a-public-open-source-ai-model-built-for-privacy/

[3] https://ethz.ch/en/news-and-events/eth-news/news/2025/09/press-release-apertus-a-fully-open-transparent-multilingual-language-model.html

[4] https://www.swiss-ai.org/apertus

[5] https://huggingface.co/collections/swiss-ai/apertus-llm-68b699e65415c231ace3b059

[6] https://publicai.co/