-
OpenAI’s agent,
ChatGPT can now perform deep research for you. It shows a summary in a sidebar with citations, and a summary that shows the process for reference.
Read the Article>
-
Gemini AI can automatically turn your spreadsheets into charts
Illustration: The Verge
Gemini has some new abilities that could make it more helpful in Sheets, Google In a blog post, Workspace announced the news. Now, Gemini will respond to questions with details about trends and static charts you can insert in your spreadsheet as images. The new feature is now available to all Workspace plans, including the $19.99 per month plan. Google One AI Premium plan.
Google claims that Gemini accomplishes all of this through the creation and execution of Python code. The code’s output is then analyzed. It may use spreadsheet formulas for simpler requests, but it will save you the headache and tedium of creating data visualizations. Gemini’s capabilities were limited before this. It could only tell you how to use Sheets and create tables.
Read the Article>
-
OpenAI CEO Sam Altman on DeepSeek R1: “an impressive model.”
The ChatGPT boss “We will deliver much better models, and it’s legitimized to have a brand new competitor,” says of his company. Then, naturally, the conversation turns to AGI.
Screenshot: @sama ( X (
-
DeepSeek says its newest AI model, Janus-Pro can outperform Stable Diffusion and DALL-E 3.
Input image analysis is limited to 384×384 resolution, but the company says the largest version, Janus-Pro-7b, beat comparable models on two AI benchmark tests.
Correction: As TechCrunch states that Janus-Pro’s image input is limited to low resolution and not its output.
-
Meta AI will use its ‘memory’ to provide better recommendations
Illustration by Nick Barclay / The Verge
Meta is widely launching the ability for its AI chatbot to “remember” certain details about you, such as your dietary preferences or your interests, the company said in a Blog post on MondayIt will then use details from your Facebook and Instagram accounts in addition to past conversations to provide more relevant suggestions.
Meta launched its AI chatbot’s memory feature last year. Now, it is available on Facebook Messenger and WhatsApp for iOS and Android users in the US and Canada. You can tell Meta AI that it should remember certain things like the fact that you enjoy traveling, but it will also “pick important details based upon context.”
DeepSeek’s top-ranked AI app is restricting sign-ups due to ‘malicious attacks’
Image: Cath Virginia / The Verge
After surging to the top of Apple’s App Store charts in the US, DeepSeek’s AI Assistant is now restricting new user sign-ups. According to a page on the incident reportregistrations are temporarily limited “due large-scale malicious attacks against DeepSeek’s service.” However, it is unclear how these restrictions are being applied.
DeepSeek’s update stated that existing users can log in normally. “Thanks for understanding and support.” A banner alert on the website said. The DeepSeek sign-up page () states that “registration is busy” but not entirely restricted. It encourages users to “try again” in the event their application fails.
Read the Article>
OpenAI has added its o1 model to Canvas.
Openai””https://x.com/OpenAI/status/1882876174533275709″”> Canvas has been added to the ChatGPT desktop application for macOS.
Character.ai responds to a wrongful death lawsuit aimed at its chatbots.
Last fall, Megan Garcia sued Character.AI, its founders, and Google over the death by suicide of her 14-year-old son, who had chatted continuously with its bots, including just before his death. In December, the firm added safety measures aimed at teens and concerns over addiction.
Google’s Gemini is already winning the next-gen assistant wars
Illustration: The Verge
One of the most important changes in Samsung’s new phones is a simple one: when you long-press the side button on your phone, instead of activating Samsung’s own Bixby assistant by default, you’ll get Google Gemini.
This is probably a good thing. Bixby was never a very good virtual assistant — Samsung originally built it primarily as a way to more simply navigate device settings, not to get information from the internet. It has gotten better since and can now do standard assistant things like performing visual searches and setting timers, but it never managed to catch up to the likes of Alexa, Google Assistant, and now, even Siri. So, if you’re a Samsung user, this is good news! Your assistant is probably better now. (And if, for some unknown reason, you really do truly love Bixby, don’t worry: there’s still an app.)
Read Article>
Microsoft opens testing for Windows AI search
Image: The Verge
Microsoft is Testing AI-powered Windows Search with a new build of the dev channel for Windows 11 Insider testers. It was announced in October and uses semantic indexing, allowing users to search for local files with more casual language. You’ll need to have a Copilot Plus PC in order to use this feature.
This feature works whether you use search boxes in Settings or File Explorer or on the taskbar. The NPU chips in Copilot Plus computers allow you to use this feature without being connected to the Internet. AI search is currently limited to Windows settings, files in image and text formats such as JPEG, PNG and PDF, TXT and XLS.
Read the Article>
“Recording is hard, so let AI do it” is a bad take.
Having lost countless nights to it, and considering my days in recording studios were some of the best of my life, Shulman seems to be either flatly lying or has no idea what he’s talking about.
Microsoft drops its GitHub Copilot Workspace waitlist.
More developers can now access Microsoft’s AI coding assistance tool that’s been on a waitlist since its debut in April last year, company CEO Satya Nadella announced in a LinkedIn post on Sunday.
Microsoft is reverting its Bing AI image generator because of quality complaints
Illustration by Haein Jeong / The Verge
Microsoft is Bing Image creator, powered by AI, is rolling back an upgrade of its model. reports TechCrunch () The rollback was prompted by weeks of complaints from users that the tool didn’t perform as well after Microsoft upgraded to a new DALL-E 3 version on December 18th.
Microsoft refused to comment on the decision to roll back things or provide specifics as to what could be causing a gap between users’ expectations and its output.
Read the Article>
Las Vegas police release ChatGPT logs from the suspect in the Cybertruck explosion
Image: LMVPD
Nearly a week after a New Year’s Day explosion in front of the Trump Hotel in Las Vegas, local law enforcement They released more information about their investigation. This includes what they know about the role that generative AI played in the incident.
The confirmed that the suspect was an active duty US Army soldier named Matthew Livelsberger. He had a “possibleManifesto” saved on the phone, along with an email to a Podcaster and other letters. They also showed video footage of him pouring fuel into the truck before driving to the Hotel, preparing for the blast. He kept a log about alleged surveillance, even though the officials claimed he had no criminal record and wasn’t being monitored or investigated.
Read the Article>
Gemini can now tell when a PDF is on your phone screen
Illustration: The Verge
In the latest version of the Files by Google app, summoning Gemini while looking at a PDF gives you the option to ask about the file, writes Android Police. According to Mishaal, you’ll need to subscribe to Gemini Advanced to use this feature. It was reported on Friday it had begun rolling out. If you have this feature, you will see a “Ask About This PDF” button when you summon Gemini and look at a PDF within the Files app. You can ask questions about the PDF by tapping the button, just as you would with ChatGPT. Google announced the screen-aware feature for the first time at its I/O developer’s conference in May.
Read the Article>
Nvidia’s $249 dev kit promises cheap, small AI power
Nvidia announced the Jetson Orin Nano Super Developer Kit, the latest in the Jetson Orin Nano AI Computer line. The tiny $249 computer is similar to a Raspberry Pi, but with powerful AI processing. It packs more AI processing punch for half the price. It’s Available to buy now
Since its introduction in 2019, the Jetson Nano line is a low-cost option for hobbyists, makers and enthusiasts to power AI and robots projects. Nvidia claims that the Nano Super’s neural processor is 70 percent faster, with 67 TOPS. This is compared to the Nano Super, which has 40 TOPS. It also has a memory bandwidth of 102GB/s which is 50 percent higher, and should speed up these operations.
Read the Article>
Dexcom adds AI reports to its OTC glucose monitor.
Dexcom’s Stelo continuous glucose monitor (CGM) for those with Type 2 diabetes is Starting to use generative AI to write weekly reports that include “more personalized tips and recommendations related to diet, sleep, and exercise” than the template previously utilized. CNBC:
Stelo AI reports do not give medical advice to users, but Dexcom is using an AI framework developed by the U.S. Food and Drug Administration [Dexcom COO Jake] Leach stated.
-
Google’s Whisk AI generator will ‘remix’ the pictures you plug in
Google has announced A new AI tool, Whiskallows you to generate images by using other images instead of a lengthy text prompt.
With Whisk you can provide images to suggest the subject, scene, and style of your AI generated image. You can also prompt Whisk with multiple pictures for each of these three things. You can also type in text prompts if you wish. You can also click the dice icon to have Google generate images for you if you don’t already have any images. (Though these images appear to be AI generated.) You can add text to a text field at the end of the search process if you wish to provide more information about the image. This is not mandatory.
Read the Article>
-
Google says the next version of its Sora competitor is better at real-world physics.
In a post announcing waitlist sign-ups for its Veo 2 video model, Google says the next version “brings an improved understanding of real-world physics and the nuances of human movement and expression.”
OpenAI’s Sora notably struggles with physics, so it will be interesting to compare the results of Veo 2 when we eventually get access.
-
Instagram’s head says social media needs more context because of AI
Illustration by Nick Barclay / The Verge
In a In a series of Threads posted this afternoon, Instagram’s Adam Mosseri said users shouldn’t believe images they see online. AI is “clearly creating” content that can be mistaken for real. He says that users should be aware of the source and that social platforms can help.
Mosseri writes that “our role as internet platforms is labeling content generated by AI as best as we can,” but he admits that “some content” may be missed by these labels. Platforms “must provide context about who’s sharing” to allow users to decide how much they trust their content.
Read the Article>
-
Listen y’all, it’s a sabotage.
Folks in the online AI research community are upset after the world’s biggest AI conference, NeurIPS, gave its prestigious Best Paper Award to, among others, a controversial former ByteDance intern named Keyu Tian, writes Wired.
ByteDance allegedly dismissed Tian for sabotaging colleagues’ AI research and hoarded resources for his own work — accusations detailed in an Anonymous GitHub blog requesting the award be revoked.
-
Searching for the first great AI app
Image: Alex Parkin / The Verge
ChatGPT launched roughly two years and two weeks ago. Now, as we near the end of 2024, the AI race is… well, where is it, exactly? It’s more competitive than ever, there’s more money being poured into new models and products than ever, and it’s not at all clear when or even whether we’re going to get products that make it all worthwhile.
On This episode of The Vergecast () discusses a variety of AI news, but all of it is based on a single theme: the tech industry’s desperate attempt to create a killer AI app. (Ideally, they would also like to see one that is profitable.) Richard Lawler, The Verge’sRichard Lawler, joins us to discuss Google Gemini 2.0 and Project Astra, Project Mariner and everything else Google does to integrate AI into the products that you use every day. We also discuss the new Android XR and Google’s renewed dedication to making headsets that work. No matter how you look, it’s an AI story.
Read article>
-
Gemini AI can now summarize what’s in your Google Drive folders
Illustration by Alex Castro / The Verge
Gemini’s integration into Google Drive is getting a little more useful. In addition to summarizing documents or answering questions about a project, the AI assistant can now Create summaries for everything in a folder
This feature allows you to open a folder, and then select the “Summarize This Folder” button at top of page. Gemini will give you a breakdown on the contents of the folder. Gemini can be used to find specific files within a folder or to ask questions, such as “What is the topic of this folder?”
Google’s AI enters its ‘agentic era’
I stepped into a room lined with bookshelves, stacked with ordinary programming and architecture texts. One shelf stood slightly askew, and behind it was a hidden room that had three TVs displaying famous artworks: Edvard Munch’s Georges Seurat, The Scream (19459503) (19659178). The Sunday AfternoonAnd Hokusai””https://en.wikipedia.org/wiki/The_Great_Wave_off_Kanagawa””>The Great Wave Off Kanagawa “There are some interesting pieces of artwork here,” said Bibo Xu. He is the lead product manager at Google DeepMind for Project Astra. Project Astra, Google’s prototype AI “universal agents”responded quickly to the question “Is there a particular piece of art that you’d like to discuss?”
It replied, “TheSaturday Afternoon art was discussed previously.” “Was there something specific you wanted to discuss or were you interested inThe Scream??”
Google, like every other company in this AI race, is frantically building AI features into its products. It’s also trying to create products that other developers will want to use. Amazon, Microsoft, Anthropic and OpenAI, on the other hand, are investing billions of dollars into solving the same problems.
This may explain why Demis Hawassabis is so excited by the new Gemini 2.0’s all-encompassing nature. He is the CEO of Google DeepMind, and head of the company’s AI initiatives. Google released Gemini 2.0 about 10 months after launching 1.5. Google still calls it an “experimental pre-release” and only one model, the smaller, lower end 2.0 Flash is being released. Hassabis insists it’s a big day.
Read the Article>
Other Stories
![Over 25 years of serving tech enthusiasts.[19459019]Record $1.5 billion crypto-heist hits exchange. Ethereum reserves drained.](https://aiobserver.co/wp-content/uploads/2025/02/3205-over-25-years-of-serving-tech-enthusiasts19459019record-15-billion-crypto-heist-hits-excha67ba8a0c5f9ce.jpg "Over 25 years of serving tech enthusiasts.[19459019]Record $1.5 billion crypto-heist hits exchange. Ethereum reserves drained.")
