Feed-aggregator

Enlarge / A woman with Long COVID who is completely bedridden, requiring the use of a wheelchair to move between rooms of her home. (credit: Getty | Rhiannon Adam)

Over four years after SARS-CoV-2's debut, researchers still struggle to understand long COVID, including the ostensibly simple question of how many people have it. Estimates for its prevalence vary widely, based on different study methods and definitions of the condition. Now, for the first time, the Centers for Disease Control and Prevention has attempted to estimate its prevalence among adults in each US state and territory. The results again show a wide range of prevalence estimates while revealing the states that were hardest hit as well as those that seem relatively spared.

Overall, the CDC found that seven states in the South, West, and Midwest had the highest prevalence of long COVID in the country, between 8.9 percent and 10.6 percent: Alabama, Montana, North Dakota, Oklahoma, Tennessee, Wyoming, and, the state with the highest prevalence of 10.6 percent, West Virginia. The results are published today in the CDC's Morbidity and Mortality Weekly Report.

On the other end of the spectrum, New England states, Washington, and Oregon had lower prevalence rates, between 3.7 percent and 5.3 percent. The lowest rate was seen in the US Virgin Islands with 1.9 percent. Washington, DC, and Guam had ranges between 1.9 percent and 3.6 percent.

Read 3 remaining paragraphs | Comments

Enlarge (credit: Getty Images)

Nvidia used to be a gaming GPU company, but the company's value has absolutely skyrocketed thanks to the AI craze and a GPU's value in accelerating AI workloads. A few purpose-built chips later, like the H100 Tensor Core GPU (with a price of $25,000-$40,000!) and A100, and Nvidia's stock is up 50 percent this year. At press time, Nvidia's market cap is now $1.8 trillion, beating Amazon ($1.76 trillion) and Google's parent company, Alphabet ($1.77 trillion), to become the world's fourth most valuable company.

How much further can Nvidia's blazing stock rally go? Next up on the 'highest market cap' list is Saudi Aramco, the national oil company of Saudi Arabia, at $2.0 trillion; then Apple, at $2.8 trillion; and another company riding the AI wave, Microsoft, at $3 trillion.

Nvidia's next earnings report is February 21. The last one was for Q3 2023, showing that the company is basically selling every AI chip it can make. Revenue was up 206 percent from the same quarter last year, and of the company's $18.12 billion in revenue, $14.51 billion was generated by its AI/data center division. Q4 will probably be another record-setting quarter for the company, and coming up in Q2 2024, Nvidia will launch its next-gen AI chip, the HGX H200 Tensor Core GPU. A TrendForce estimate puts Nvidia's AI server market share at 60–70 percent.

Read 3 remaining paragraphs | Comments

Enlarge (credit: Aurich Lawson | Getty Images)

An open source project that let people view tweets without going to Twitter.com has shut down, saying that Elon Musk's changes closed off all possible ways to access the Twitter network without a user account.

Nitter provided an alternative front-end to Twitter but has been struggling for months. The website Nitter.net went down a few weeks ago, and the project announced its demise today. "Nitter is over—it's been a fun ride. Twitter blocked the last known way to access their network without a user account," the update said.

There were multiple instances of Nitter. Users reported that many instances went down about eight months ago as Twitter (now called X) imposed new API restrictions. Some instances stayed online with workarounds, which no longer work.

Read 12 remaining paragraphs | Comments

There's a new trailer for Godzilla x Kong: The New Empire, coming to theaters next month.

Warner Bros. has released a new trailer for Godzilla x Kong: The New Empire, directed by Adam Wingard. It's the fifth feature film in the rebooted franchise, which also includes the animated series Skull Island and Apple TV+'s Monarch: Legacy of Monsters.

(Spoilers for Godzilla vs. Kong below.)

As previously reported, Godzilla x Kong picks up sometime after its 2021 predecessor. Godzilla vs. Kong showcased not only a major showdown between its titular titans—in which Godzilla emerged the victor—but also the two teaming up in the climactic finale to take out Mechagodzilla, a telepathically controlled creature with the severed head of Ghidorah. Ghidorah's consciousness took over when Mechagodzilla was activated, and it took both Kong and Godzilla (plus some timely help from humans) to defeat him. (Kong got the final honors, although Godzilla charged the killing ax—made from one of his dorsal plates—with his atomic breath.)

Read 5 remaining paragraphs | Comments

Enlarge (credit: Nvidia)

On Tuesday, Nvidia released Chat With RTX, a free personalized AI chatbot similar to ChatGPT that can run locally on a PC with an Nvidia RTX graphics card. It uses Mistral or Llama open-weights LLMs and can search through local files and answer questions about them.

Chat With RTX works on Windows PCs equipped with NVIDIA GeForce RTX 30 or 40 Series GPUs with at least 8GB of VRAM. It uses a combination of retrieval-augmented generation (RAG), NVIDIA TensorRT-LLM software, and RTX acceleration to enable generative AI capabilities directly on users' devices. This setup allows for conversations with the AI model using local files as a dataset.

"Users can quickly, easily connect local files on a PC as a dataset to an open-source large language model like Mistral or Llama 2, enabling queries for quick, contextually relevant answers," writes Nvidia in a promotional blog post.

A screenshot of Chat With RTX, which runs in a web browser window. (credit: Benj Edwards)

Using Chat With RTX, users can talk about various subjects or ask the AI model to summarize or analyze data, similar to how one might interact with ChatGPT. In particular, the Mistal-7B model has built-in conditioning to avoid certain sensitive topics (like sex and violence, of course), but users could presumably somehow plug in an uncensored AI model and discuss forbidden topics without the paternalism inherent in the censored models.

Also, the application supports a variety of file formats, including .TXT, .PDF, .DOCX, and .XML. Users can direct the tool to browse specific folders, which Chat With RTX then scans to answer queries quickly. It even allows for the incorporation of information from YouTube videos and playlists, offering a way to include external content in its database of knowledge (in the form of embeddings) without requiring an Internet connection to process queries.

We downloaded and ran Chat With RTX to test it out. The download file is huge, at around 35 gigabytes, owing to the Mistral and Llama LLM weights files being included in the distribution. ("Weights" are the actual neural network files containing the values that represent data learned during the AI training process.) When installing, Chat With RTX downloads even more files, and it executes in a console window using Python with an interface that pops up in a web browser window.

Several times during our tests on an RTX 3060 with 12GB of VRAM, Chat With RTX crashed. Like open source LLM interfaces, Chat With RTX is a mess of layered dependencies, relying on Python, CUDA, TensorRT, and others. Nvidia hasn't cracked the code for making the installation sleek and non-brittle. It's a rough-around-the-edges solution that feels very much like an Nvidia skin over other local LLM interfaces (such as GPT4ALL). Even so, it's notable that this capability is officially coming directly from Nvidia.

On the bright side (a massive bright side), local processing capability emphasizes user privacy, as sensitive data does not need to be transmitted to cloud-based services (such as with ChatGPT). Using Mistral 7B feels similarly capable to early 2022-era GPT-3, which is still remarkable for a local LLM running on a consumer GPU. It's not a true ChatGPT replacement yet, and it can't touch GPT-4 Turbo or Google Gemini Pro/Ultra in processing capability.

Nvidia GPU owners can download Chat With RTX for free on the Nvidia website.

Read on Ars Technica | Comments

Enlarge / Musician Benn Jordan, who performs under the alias The Flashbulb, successfully defended his music against streaming fraud allegations.

Last Friday, musician Benn Jordan assumed his phone was glitching when he tried to pull up one of his albums and couldn't find it on Spotify. Then he noticed all the notifications he'd gotten from fans asking why he'd removed his music on all the streaming platforms where his music could typically be found, including Apple Music, iTunes, Deezer, and YouTube Music.

But Jordan had not made any such decision. By the time night fell on Friday, the gravity of what had happened finally sank in, and he realized something was "very, very wrong."

For the past 17 years, Jordan has paid his digital distributor, TuneCore, thousands of dollars to manage his music on streaming platforms. Under his alias The Flashbulb, Jordan had released more than a dozen albums, reaching 1.9 million listeners on Spotify who added his songs to more than 300,000 playlists last year alone. In total, he had earned over $400,000 in sales for TuneCore since signing up for its services in 2007.

Read 28 remaining paragraphs | Comments

Dat blijkt uit de uitspraak die het Hof hierover op 15 februari deed. De NVJ en het Persvrijheidsfonds vorderden dat zij zich in de zaak mochten voegen aan de kant van DPG, de uitgever van het AD. NVJ wil in de zaak opkomen voor de belangen van journalisten in het algemeen, en voor haar leden. Ook het Persvrijheidsfonds komt op voor zijn achterban van journalisten.

Het hof heeft dit nu toegestaan. Ze beslist dat de twee partijen voldoen aan de ontvankelijkheidseisen voor een collectieve actie in kort geding. Een door de Orde van Advocaten genoemde omstandigheid dat de hoofdzaak geen collectieve actie is, staat niet in de weg aan toelating van NVJ en Persvrijheidsfonds als aan de zijde van DPG gevoegde partij.

Advocaat Royce de Vries, de zoon van de vermoorde misdaadjournalist Peter R. de Vries, spande het kort geding aan vanwege een geplande, verdere publicatie door het AD op basis van geluidsopnames. Vermoed wordt dat de opnames zijn gemaakt door zijn vader. Daarop verbood de Amsterdamse rechtbank het AD de publicatie van een nieuw verhaal. De voorzieningenrechter concludeerde enkele weken terug in zijn uitspraak in kort geding dat de belangen van de eiser en de orde zwaarder wegen dan het belang van de persvrijheid. Daarmee bedoelde ze ‘de bescherming van de rechten van anderen’ en het ‘voorkomen van verspreiden van vertrouwelijke mededelingen’.

Eerder had het AD al een verhaal gepubliceerd naar aanleiding van de opnames, over mogelijke omkoping van een justitiemedewerker door ex-strafrechtadvocaat en de inmiddels door deze affaire als BNNVARA-presentator gestopte Khalid Kasem. De voormalige kantoorgenoot van Peter R. de Vries ontkent de beschuldigingen die via het artikel zijn geuit.

Het gerechtshof in Amsterdam meldt dat het spoedappèl in deze zaak zal worden behandeld op 19 februari.

Het bericht Journalistenvakbond mag zich voegen in hoger beroep Royce de Vries / AD verscheen eerst op Mr. Online.

Enlarge (credit: iStock via Getty Images)

You shouldn’t trust any answers a chatbot sends you. And you probably shouldn’t trust it with your personal information either. That’s especially true for “AI girlfriends” or “AI boyfriends,” according to new research.

An analysis into 11 so-called romance and companion chatbots, published on Wednesday by the Mozilla Foundation, has found a litany of security and privacy concerns with the bots. Collectively, the apps, which have been downloaded more than 100 million times on Android devices, gather huge amounts of people’s data; use trackers that send information to Google, Facebook, and companies in Russia and China; allow users to use weak passwords; and lack transparency about their ownership and the AI models that power them.

Since OpenAI unleashed ChatGPT on the world in November 2022, developers have raced to deploy large language models and create chatbots that people can interact with and pay to subscribe to. The Mozilla research provides a glimpse into how this gold rush may have neglected people’s privacy, and into tensions between emerging technologies and how they gather and use data. It also indicates how people’s chat messages could be abused by hackers.

Read 15 remaining paragraphs | Comments

Enlarge / President Vladimir Putin visiting the Rocket and Space Corporation (RSC) Energia in Korolyov, outside Moscow, on October 26, 2023. (credit: GRIGORY SYSOYEV/POOL/AFP via Getty Images)

When you think about the prospect of Russia putting nukes in space, you can understand why some people might fill in the blanks and envision a surprise apocalyptic bombardment from orbit.

That is the nightmarish prophecy members of Congress tried to extinguish Wednesday after news leaked that US intelligence officials obtained information about Russian preparations for warfare on the high frontier.

It all began with a cryptic statement Wednesday morning from Rep. Mike Turner (R-Ohio), chairman of the House Intelligence Committee. Turner referenced intelligence information that the committee would begin sharing with all members of Congress about a "serious national security threat."

Read 24 remaining paragraphs | Comments