Huang’s most interesting offering at the show was the new Groq 3 LPX, a custom rack made for inference loads.
Continue reading “Nvidia will sell $1 trillion of its AI chips by 2027, launches inference rack”
After pivoting hard to AI content, BuzzFeed faces questions of survival
Back then, it was hiring a new staffer every day, and developed a Pulitzer Prize winning newsroom.
Fast forward to January 2023 and CEO Jonah Peretti said the company would pivot hard to AI quizzes, and in April the same year, he shut down the newsroom and said that AI would replace most of the static content on the site, Futurism reports.
Three years later, the company’s quizzes didn’t take off and the AI content was underwhelming — leaving the company with a net loss of $57.3 million in 2025.
Continue reading “After pivoting hard to AI content, BuzzFeed faces questions of survival”
Entering the halls of power, AI chatbots get approved for U.S. Senate use
The chatbots cleared for «drafting and editing documents, summarizing information, preparing talking points and briefing material, and conducting research and analysis» are ChatGPT Enterprise, Gemini on Workspace and Copilot, according to Business Insider.
Anthropic’s Claude is not on the list, although it has been approved for use in The House of Representatives since September 2024. President Trump earlier said government agencies shouldn’t use Claude after the Pentagon spat in February.
The Sergeant at Arms will now provide all Senate employees with an AI chatbot license at no cost. The office also touts Copilot Chat as integrated with Office 365 and is usable with Word and Excel — although Claude also offers this capability.
Read more: Business Insider has the memo, NY Times reported it first. Popvox has a Congressional AI tracker.
Internet legend Digg’s relaunch on hold after deluge of AI slop
— Building on the internet in 2026 is different. We learned that the hard way, writes CEO Justin Mezzell, — We knew bots were part of the landscape, but we didn’t appreciate the scale, sophistication, or speed at which they’d find us. We banned tens of thousands of accounts. We deployed internal tooling and industry-standard external vendors. None of it was enough.
Launched in 2004, as some readers might remember, Digg was once not just viral, but a virality machine. It was one of the original link sharing and commenting platforms that could see massive engagement and send great amounts of traffic to pages getting upvoted.
Continue reading “Internet legend Digg’s relaunch on hold after deluge of AI slop”
Meta postpones launch of «Avocado» model until «at least» May
Pushing back the launch from «early 2026» is due to performance issues, The New York Times reports — as the model «currently falls between Gemini 2.5 and Gemini 3.»
This falls far behind the Meta Superintelligence Lab’s ambitions of «pushing the frontier,» after working on the new model «for months,» Reuters reports.
— We’re excited for people to see what we’ve been cooking very soon, a Meta spokesperson told Reuters.
Apparently, Meta AI has considered «temporarily» licensing Google’s Gemini model to propel their AI products while they wait for their own models to catch up.
Read more: The New York Times (paywalled) and Reuters.
«Ask Maps» brings Gemini 3 intelligence, personalization to Google Maps
The feature is supposed to work great for questions of where to find the nearest restroom, or a cozy vegan restaurant nearby — and it even lets you book a table right from the app.
To achieve this, Gemini will scan information from the Maps database consisting of some 300 million places and reviews from over 500 million contributors to find you just the right spot.
Ask Maps also remembers your previous saved spots or queries, so it will know that you are vegetarian, say, or if you have any special needs or preferences.
Of course, once you find a spot, Maps will help you navigate to get there — and in the biggest update in a decade, you now get a 3D driving experience.
Ask Maps is only available on mobile in the USA and India, with desktop support «coming soon.»
Read more: Google’s announcement, The Verge, Engadget.
Anthropic introduces charts and diagrams in Claude, days after ChatGPT
Claude can now build interactive charts and diagrams, directly in the chat.
Available today in beta on all plans, including free.
Try it out: https://t.co/tHPAZRgQkn pic.twitter.com/WXRrD4VkAt
— Claude (@claudeai) March 12, 2026
Claude can now illustrate some concepts and processes within the main chat window, just days after ChatGPT added visuals for some math queries.
Previously, Claude could draw illustrations in a sidebar window that you could copy or download, but these can be interactive and are made inside the main chat, writes The Verge.
Sometimes, the chatbot will determine itself if a concept needs illustrating, or you can simply ask it to make one yourself — and it will draw a chart from html and xml vectors.
The feature is available to all users, paid and free — but it’s officially in beta, so users can expect hiccups, and it’s not available on mobile, notes Engadget.
Gemini on Chrome expands to more countries and languages
Gemini in Chrome adds a new side panel, letting you chat with Gemini without opening up a new tab, and can do things like summarize or interact with web pages. It can connect to Gmail, Calendar, YouTube, Shopping and flights information.
It also comes with Nano Banana features, so you can try an apartment listing picture with your own furniture, for example.
In addition to the three new countries, which are mostly English-speaking, Google is announcing support for another 50 languages.
This of course includes Hindi, but there is also support for French, Spanish, Chinese and lots of other European languages.
Read more: Google’s announcement. Writeups on TechCrunch and Engadget.
Sora 2 video generation could be heading for ChatGPT «soon»
According to The Information’s x.com post, the release seems imminent, while Blaho has spotted Sora generation mentions and ID’s in the latest Andoid beta ChatGPT App.
Sora 2 was launched as a quasi social network with its own app in October 2025 and was invite-only.
Judging by early search interest on Google for «Sora invite code» soon after launch, interest for the app greatly exceeded capacity and available invites.
It produces 10-second clips from prompts or pictures and promises to put your likeness in anything you can imagine. It briefly went viral, running up some 920 million weekly users before dropping from #1 to #165 in the App Store.
Putting the app’s features into the mainline GPT app will bring it up to par with Google’s AI subscriptions, which gives users access to the video generator Veo 3 on the $25 Pro plan.
Read more: The Information (paywalled), Tibor Blaho on X, and Reuters.
AI lab Thinking Machines gets investment, gigawatt compute from Nvidia
The startup has entered into «a multiyear strategic partnership» that will provide them with both money and significant Vera Rubin compute early next year — about the same level the first version of Grok was trained on.
The parties are not disclosing a sum total, but 1 GW of Nvidia compute is estimated to be worth about $50 billion, Reuters notes.
Murati’s AI lab has been largely secretive about its actual products, releasing a configurable API in December 2025 and vowing to make AI models more accessible, capable and, yes, customizable.
They raised $2 billion at a $12 billion valuation from Andreessen Horowitz and Nvidia in July 2025.
Read more: Joint press release, writeups on Reuters, CNBC and TechCrunch.
Meta announces new processor generation focused on inference
The new generation Meta Training and Inference Accelerator (MTIA) isn’t built to replace the chips sourced from AMD or Nvidia, but are intended to supplement them and achieve «the lowest possible price.»
MTIA-chips are taking a different tack on developing AI models, optimizing for inference — the process of answering AI queries — instead of training. Most chips are customized for training, which is more compute intensive, but the majority of a chip’s life is spent putting together answers in production.
— Chip designs are based on projected workloads, but by the time the hardware reaches production — often two years later — those workloads may have shifted substantially, Meta says in their press release.
The new chips «have either already been deployed or are scheduled for deployment in 2026 or 2027,» Meta says — and they don’t disclose just how many of these they are making.
Read more: Meta’s press release. Writeups on Reuters and CNBC.
Claude for Excel and Powerpoint now shares info, and gets skills
That eases some hassle for those who do a lot of work in PowerPoint and Excel, and removes the need to reintroduce the task or use extra steps.
It means that users can pull financials into a workbook and drop the valuation summary into a PowerPoint slide without switching tabs or re-explaining at every step.
At the same time, Anthropic is launching skills for workflows — that can be shared and dropped into other apps in an organization, so everyone can use the same time-saving actions stored in them.
Skills are stored prompts for workflows, and work as an old-school template; with everything set and working on repetitive tasks that are easy to automate.
Read more: Anthropic’s presentation and launch tweet. Writeups on VentureBeat and The Decoder.
As 140 million people use ChatGPT for learning math, OpenAI adds visuals
140 million people use ChatGPT to help them understand math and science concepts every week. pic.twitter.com/hIK2t7joKN
— ChatGPT (@ChatGPTapp) March 10, 2026
That’s a big number out of 900 million weekly ChatGPT users, accounting for about a sixth of the total usage and outperforming even health questions.
In their announcement, OpenAI says that more than half of U.S. adults struggle with math, and don’t feel confident teaching it to their kids.
Continue reading “As 140 million people use ChatGPT for learning math, OpenAI adds visuals”
French Canal+ partners with Gemini to index content, ease production
Starting in June 2026, Canal+ will be providing producers with access to Veo 3, which will help creators «pre-visualize scenes before shooting,» or help recreate history from a single image.
They will also be indexing its extensive content library to use in a Gemini-powered recommendation engine, creating a massive, multimodal database of sound, video, and text data.
Canal+ recently bought the pan-African boradcaster MultiChoice, extending the reach of its app and services to some 70 countries, as they are racing toward a target of 100 million streaming users, and hope to compete with the American streaming giants.
Read more: Canal+ press release, Reuters report.
Ex-Meta chief AI scientist raises $1 billion to build world models
CEO Alexandre LeBrun notes to TechCrunch that world models are the latest buzzword in AI, but cautions that the startup is doing «fundamental research» and won’t be launching any products «for years.»
LeCun and LeBrun both say that generative AI models work by predicting text or images, and are severely limited when it comes to interacting with the real, messy, and unpredictable world.
AMI sees their clients as being in «industrial process control, automation, wearable devices, robotics, healthcare, and beyond,» according to their front page.
According to Reuters, their technology might show up in Meta’s AI glasses «in the short term.»
Read more: Reuters, TechCrunch and Wired.