Microsoft’s Bing Chat is broadening its horizon and starting testing on Google Chrome and Apple’s Safari


’Microsoft’s Bing Chat comes to Chrome and Safari in tests for ‘select users

Microsoft’s chatty AI buddy, Bing Chat, is spreading its wings and trying out life on other web browsers, not just Microsoft-owned ones. It’s making its debut on Google Chrome and Apple’s Safari, opening up to a bigger crowd. Before this, Bing Chat was only hobnobbing with folks using Microsoft products, like the Bing mobile app and the Edge browser.

Microsoft didn’t officially announce this, but they confirmed it when TechCrunch asked about it. They’re still testing the waters and it’s only available to a select bunch right now. Bing Chat has some neat new features up its sleeve. If you’re running Windows 10 or 11, you might get a pop-up on your taskbar offering you a chance to try out Bing Chat on Chrome.

But it’s not all smooth sailing. Folks testing the new AI chat on other browsers have run into some speed bumps. It seems like Bing Chat on Chrome is a bit stingy, only allowing five messages per conversation, compared to the generous 30 on Microsoft Edge. Plus, it caps the character count to 2,000, instead of the whopping 3,000 on Edge.

Microsoft’s keeping mum about the details. We don’t know when the expansion started, which platforms are supported, or if they’re going to roll it out globally. They’re also reportedly testing a native dark theme, but it’s not available to everyone yet. Microsoft’s been busy pushing Bing Chat into more of its products, and it’s even going corporate with features that cater to business data privacy. Plus, it’s testing a Visual Search feature, where the chatbot answers questions about uploaded images. We’ll just have to wait and see how it all plays out.


SoftBank forms JV to build AI-powered warehouses with Symbotic

Japanese tech giant SoftBank is partnering up with supply-chain whiz Symbotic to build fancy warehouses run by artificial intelligence. Their new venture, GreenBox Systems, is going to be majorly owned by SoftBank. Together they’re pumping $100 million into this gig. GreenBox is also buying a boatload of AI tech from Symbotic, worth around $7.5 billion.

This deal sweetens the pot for SoftBank, already a Symbotic backer, by handing over warrants equal to about 2% of the U.S. firm’s outstanding shares. They also bought a big chunk of Symbotic stock from the CEO, Rick Cohen.

The news sent Symbotic’s stock climbing 1.3% and SoftBank’s up 1.2% in Tokyo trading. This move shows SoftBank’s all-in on AI, switching gears from defense to offense, as CEO Masayoshi Son puts it. The success of AI tech like ChatGPT this year has investors throwing cash at anything AI, even as overall funding is getting thin due to a shaky economy.

SoftBank’s exec, Vikas Parekh, believes GreenBox will unlock the massive potential of AI in supply chains. SoftBank will hold a 65% stake in GreenBox, with Symbotic owning the rest. Starting in 2024, GreenBox will begin ordering Symbotic’s systems over six years.


Meet FreeWilly, Large And Mighty Instruction Fine-Tuned Models

Stability AI and its CarperAI lab are showing off their new language models, FreeWilly1 and FreeWilly2. These high-powered models are open for anyone to use, and are good at reasoning and other complex tasks. FreeWilly1 and FreeWilly2 are based on different foundation models and fine-tuned with a unique dataset. They’re pretty much research experiments and are out there for non-commercial use.

Stability AI trained these models in a way similar to how Microsoft did it in one of their research papers. The data used to train FreeWilly came from datasets created by Enrico Shippole, and it amounts to about 600,000 data points. That’s only 10% of what Microsoft used for their model, but FreeWilly still shows off some impressive skills.

When they tested these models internally, they found FreeWilly excelled in a bunch of areas, like understanding tricky language and answering complex questions. On July 21st, 2023, researchers at Stability AI, and then again independently by Hugging Face, tested FreeWilly using an AI evaluation system and posted the results online.


Google teams up with climate scientists to monitor permafrost with AI

Google’s philanthropic branch,, is teaming up with scientists to develop a first-of-its-kind tool for keeping an eye on thawing permafrost in the Arctic in near real-time. Given the Arctic is heating up three to four times faster than the rest of the world, this thawing could release massive amounts of greenhouse gases, making this tool pretty important.

Up to now, scientists have had to physically go to sites in the Arctic to keep tabs on permafrost. But and the Woodwell Climate Research Center are working to change this. With a $5 million grant, the two are creating a new satellite-based method to monitor permafrost, using AI and other new techniques.

This online tool, which doesn’t currently exist, would provide real-time, high-res monitoring of permafrost regions and be freely available to everyone. This tool could help fill some gaps left by geopolitical tensions, like scientists not being able to work with Russian collaborators due to Russia’s invasion of Ukraine.

The project aims to allow scientists to keep track of permafrost thaw on a monthly basis and estimate carbon loss. Satellite images would come from public platforms and possibly commercial imagery from companies like Planet and Maxar. However, Anna Liljedahl, a Woodwell climate scientist, points out that this doesn’t mean we can forget about fieldwork. Both remote sensing and fieldwork need to come together for a complete picture.


The NYC subway is using automated scanning software to spot fare evaders

NYC subways are getting smart to fare dodgers. The city’s transit honchos are trialing AI surveillance tech in some stations to nab folks skipping out on paying. They’re tight-lipped on the exact locations, but we know 24 more stations will join the game by year-end.

The transit authority’s out nearly $300 million thanks to fare skippers last year. They’re using a software whipped up by Spanish AI dudes at AWAAIT. This smart cam sends pics of fare cheaters straight to subway staff’s phones. The data’s showing trends too, like peak skipping time is when the school bell rings, and most folks sneak through emergency gates. The vids don’t hang around forever though – they’re kept for just a short while.


Nvidia Jumps Higher as Mizuho Analysts See $300 Billion AI Chip Potential

Nvidia shares jumped higher after analysts at Mizuho increased their price target for the tech group, citing the potential of Nvidia’s new focus on AI chips to generate around $300 billion in additional revenue. According to analyst Vijay Rakesh, the overall market for AI chips could increase tenfold over the next five years to more than $400 billion. Nvidia could command a 75% market share, even with rising competition from server groups such as Amazon, Google, and Tesla. This could lead to approximately $300 billion in AI-specific revenue.

Rakesh further suggested that Nvidia stock could rise to as high as $769 per share, corresponding to a market value of $1.9 trillion based on the most optimistic revenue gain assessment. His mid-point price estimate is $625 per share by 2027, and he raised his current Nvidia price target by $130, to $530 per share.

Nvidia shares were marked 1% higher in early Monday trading, extending the stock’s year-to-date gain to around 207%. The company joined the ‘Trillion Dollar Club’ last month after reporting impressive first-quarter earnings. Only five U.S. companies have exceeded the $1 trillion market cap threshold. Last month, Nvidia predicted current-quarter revenue of around $11 billion, more than 50% ahead of Wall Street forecasts.


Intel Accelerates AI Development with Reference Kits

Intel’s going all out to turbocharge AI development. They’ve rolled out 34 open source AI reference kits. These kits offer everything from model code to training data and libraries. They’re aimed at making AI development quicker and easier for developers and data scientists, whether they’re working in-house, up in the cloud, or on the edge.

The idea? Streamline the whole AI process, beef up existing smarts, and get solutions deployed lickety-split. These kits span industries like consumer goods, healthcare, finance, and more. They’re showing real results too: chatbots responding 45% quicker, defect detection in life sciences 55% faster, and predicting the health of utility assets with 25% more accuracy.


AI talent is in high demand even at non-tech companies. And jobs are paying way over six figures — up to $300,000.

Artificial Intelligence (AI) skills are in high demand across various industries, including those outside the traditional tech sector, with salaries reaching up to $300,000. Companies from various fields such as law, insurance, and agriculture are looking for AI talent to leverage their in-house data more efficiently and make improved predictions and decisions.

The demand for professionals with knowledge in AI research, machine learning, and deep learning is surpassing the available supply, leading to a sharp increase in salaries. The base salary ranges for AI researcher roles at non-tech firms can range from $150,000 to $250,000.

The demand for AI skills underscores a broader strategy where non-tech firms are licensing commercially available AI technology and trying to develop their proprietary tools. As AI becomes increasingly mainstream, professionals in all industries will need to adapt and acquire the relevant skills.


Top News Publishers Are Reportedly Planning To Sue AI Firms

Several top publishers, including The New York Times, News Corp, Axel Springer, and Dotdash Meredith owner IAC, are reportedly forming a coalition to sue large AI firms, such as Google and OpenAI, for using their content to train generative AI models without their permission. This news comes amid concerns about how AI could impact website traffic and the potential loss of attribution and linking when AI chatbots scrape data from publisher pages.

IAC CEO Joey Levin has voiced concerns about the profound impact AI might have on the news media industry. This initiative follows warnings from IAC Chairman Barry Diller about AI’s potentially “catastrophic” effect on publishing.

This legal action is contrary to the recent agreement between the Associated Press (AP) and OpenAI to license an archive of news stories, which will enable AP to tap into OpenAI’s “technology and product expertise.”

This dispute occurs within the broader context of numerous lawsuits against AI companies such as Google, OpenAI, Meta, and Midjourney, who have been accused of training their models on content without consent or compensation.