❌

Vue normale

Il y a de nouveaux articles disponibles, cliquez pour rafraîchir la page.
À partir d’avant-hierFlux principal

Neurapix: a new AI start-up from Germany that learns from previously edited images and applies individual’s editing style to new photos in Adobe Lightroom Classic

✇Photo Rumors
Par : PR admin
26 février 2024 à 16:42


Neurapix is a new start-up from Goettingen in Germany that has developed an artificial intelligence (AI) known as the SmartPreset. It learns from previously edited images and can apply the individual's editing style to new photos in Adobe Lightroom Classic at lightning speed. This is particularly beneficial for photographers who frequently shoot large events such as weddings, as it allows for significant time savings.

Neurapix recently released their plugin for Adobe Lightroom Classic.

PhotoRumours readers will get 2,000 free image edits if they register via that link.


Additional information:

General

The SmartPreset, as opposed to a traditional static preset for Lightroom, is based on artificial intelligence. This allows the SmartPreset to understand and learn from a photographer's individual editing style through his/her past edits. As a result, it can consistently apply the learned style to future images in Lightroom, especially in various lighting situations. This capability was previously only achievable through tedious manual slider adjustments. The SmartPreset is able to quickly set all necessary sliders for a specific editing style, resulting in a significant time savings of around 90% for photographers. Despite this automation, photographers still maintain full control over the editing process and have the ability to fine-tune the slider values as needed for their RAW files. Additionally, the SmartPreset regularly analyzes post-processing corrections and continually improves its understanding of how final images should appear.

In mid-January, the Gottingen-based company announced that it can now process 1,000 images per minute using a SmartPreset.

PhotoRumours readers will get 2,000 free image edits if they register via that link. After that, the cost is $0.03 per photo. For example, the editing of 500 images from a wedding would typically cost only $15.

To get there is very easy: After installing the Neurapix plugin, the photographer selects all photos from an unedited photo event in Adobe Lightroom Classic and click on “Create SmartPreset” via the menu (Library > Plug-in Extras) and selects “Kickstart”. The Neurapix algorithm then automatically selects 20 representative images for the event, encompassing a wide range of lighting situations and settings. In the background, all event photos are already being uploaded to the secured Neurapix server.

The photographer now edits the 20 representative photos in his/her own style and then clicks on "Kickstart" again. After naming their style (the future name of the SmartPreset), the 20 edited images are uploaded. On the Neurapix server, all event photos are then edited in the newly learned style and ready for download in just a few minutes. After potential manual corrections, the photo event is ready for use.

Reduce your photo editing time in Lightroom by 90%

Do you want to shoot more and edit less? Spend more time with your family, friends, and the things you really like? Or maybe expand your photo business? Then this AI plugin can revolutionize your workflow. The more pictures you have to edit and deliver in a short time for your clients, the more valuable it gets!

German company Neurapix has developed artificial intelligence that is able to learn your individual image editing style from previously edited images and apply it within Adobe Lightroom Classic. This allows you to have large numbers of photos edited in your own style at lightning speed, resulting in significant time savings of 90%.

One of the things photographers like most about Neurapix: it is seamlessly integrated into their classic workflow. You don't have to switch to another software but only install the Neurapix plugin via the Adobe Store once you have opened your account. After that, every single step happens within Adobe Lightroom Classic. The only difference to your old workflow: you select a dynamic SmartPreset instead of your static preset.

Consistency is key: Why SmartPresets have an edge over normal presets

The SmartPreset, as opposed to a traditional preset for Lightroom Classic, is based on artificial intelligence. This allows it to understand the situation and conditions the image was taken in. As a result, it can apply a very consistent look in your individually trained style - especially in various lighting situations! Previously, this was only achievable through tedious manual slider adjustments.

Despite this automation, you still maintain full control over the editing process and have the ability to fine-tune the slider values as needed for your RAW files. Manual adjustments are relatively rare though.

Train your AI with only 20 edited images!

Neurapix has just released a new AI training option called "Kickstart" which allows you to get your own SmartPreset even faster in only a few minutes. After installing the Neurapix plugin, you upload all photos from an event that you want to edit through the "Kickstart" feature. An algorithm analyzes the event and selects 20 representative images, encompassing a wide range of different lighting situations and settings.

You then edit those 20 photos in the style you want (you can also easily create unlimited new SmartPresets very quickly this way). Neurapix will analyze your style and apply it to all other pictures of your photo event. You will receive your results in just a few minutes. After potential manual corrections, your event is ready for use and with one more click, you can finalize your SmartPreset.

The more often you use your SmartPreset, the better it will get in understanding your editing style. Feel free to give feedback on the results and your AI will be re-trained regularly.

Fair pricing and a maximum of flexibility

The AI training to create your personal SmartPreset is always free, as are regular improvements made to it by the Neurapix AI. Additionally, each user is provided with three free default SmartPresets for direct use. More SmartPresets from top photographers can be found in the Neurapix Store. The first 1,000 edits (regular test phase) are always free as well.

Get 2,000 FREE EDITS from PhotoRumours to try Neurapix

After the test period, you can choose between two pricing options. With the "Pay-per-Picture" model you only pay for the photos you edit ($0.03 per picture for the first 1,000 per month, afterward only $0.02). If you shoot a lot of events, the "AI Flat Rate" might be the even better choice for you: Get unlimited edits and your photos in real-time for a capped amount of money (annual deal for $49.95 per month, or monthly rate of $79.95 which can be canceled at any time). Processing will be done locally on your own PC, if you decide on this option.

Neurapix also offers automatic and complimentary straightening of all photos, eliminating the need for manual adjustment of pictures. The new straightening function saves you even more time and comes at $0.01 per photo if you're using the Pay-per-Picture option. In the AI Flat Rate, straightening is included.

Real-time editing and data security guaranteed

You will be impressed by Neurapix's lightning speed: If you're using the Pay-per-Picture option, your images will be uploaded to the company's servers and be edited with a speed of 600 images per minute! By the way, data security is a top priority for Neurapix: All photos transferred for AI training and editing are processed exclusively on Neurapix's own servers in Germany, ensuring the security of all client data. All processes are fully compliant with the General Data Protection Regulation (GDPR).

Users of the "AI Flat Rate" get their edited images even faster because all processing is done locally on the photographer's computer so they are independent of upload and download times as well as server capacities. After 20 to 30 seconds, the first few pictures are edited already and you can start your final review right away while the rest of your photos is continuing to be edited.

"At Neurapix, we want to make life easier for photographers," says Simon Diegmann, co-founder and CEO of Neurapix. "We know from them that editing events is very a time-consuming and often tedious task. Neurapix is like a personal editing assistant, but better, faster, and available 24:7. Your SmartPreset delivers your edited photos in exactly the style you want. Also, we're giving photographers a maximum of flexibility offering the different pricing options so they can choose what fits best for them."

Create an account, download the Neurapix plugin, and try it for free: neurapix.com.

The post Neurapix: a new AI start-up from Germany that learns from previously edited images and applies individual’s editing style to new photos in Adobe Lightroom Classic appeared first on Photo Rumors.

Topaz Video AI 4.2 coming with several new features

✇Photo Rumors
Par : PR admin
27 février 2024 à 00:32


Topaz Labs will soon release Video AI 4.2 with several new features:

NEW: Plugin release for Davinci Resolve

  • Video AI available in Davinci Resolve as an OpenFX plugin
  • All Enhancement models + Motion Deblur
  • Available in both the Edit and Color workspaces

NEW: Aion Frame Interpolation model

  • More accurate motion estimation
  • Reduced artifacting and ghosting
  • True 16x slow motion
  • Improved temporal consistency
  • Reduced tiling artifacts

3D LUT Export

  • Process log footage directly in Video AI

Topaz Photo AI version 2.3.0 released, 2024 roadmap revealed

Topaz Gigapixel 7 released

The post Topaz Video AI 4.2 coming with several new features appeared first on Photo Rumors.

Elon Musk porte plainte contre OpenAI, le créateur de ChatGPT

✇Numerama.com - Magazine
Par : Nicolas Lellouche
1 mars 2024 à 11:17

Très en colère contre OpenAI, à qui il reproche d'avoir trahi sa mission initiale, Elon Musk attaque l'entreprise et son fondateur. Le milliardaire, qui a depuis lancé son propre concurrent d'OpenAI, fait tout pour nuire au créateur de ChatGPT.

Copilot For OneDrive Will Fetch Your Files and Summarize Them

✇Slashdot
Par : BeauHD
2 mars 2024 à 10:00
An upcoming April release of Copilot for OneDrive will be able to find, summarize, and extract information from a wide range of files, including text documents, presentations, spreadsheets, HTML pages and PDF files. "Users can ask Copilot to tailor summaries to their liking, such as only including key points or highlights from a specific section," reports The Verge. From the report: The chatbot will also be able to respond to natural language prompts and answer highly specific questions about the contents of a user's files. Some examples given by Microsoft included asking Copilot to tabulate a week's worth of beverage sales and throw the data in a table view by day. Or, asking it to list the pros and cons of a project, or display the most recent or relevant files. Users can even ask Copilot for advice on how to make their documents better. Copilot on OneDrive will also be able to create outlines, tables, and lists for users, based on existing files. A few examples given were: - Using the /sales-enablement.docx as reference, create an outline of a sales pitch to a new customer. - For these selected resumes, create a table with names, current title, years of experience, educational qualifications, and current location. - Create a list of frequently asked questions about project Moonshot.

Read more of this story at Slashdot.

How AI is Taking Water From the Desert

✇Slashdot
Par : EditorDavid
3 mars 2024 à 19:03
Microsoft built two datacenters west of Phoenix, with plans for seven more (serving, among other companies, OpenAI). "Microsoft has been adding data centers at a stupendous rate, spending more than $10 billion on cloud-computing capacity in every quarter of late," writes the Atlantic. "One semiconductor analyst called this "the largest infrastructure buildout that humanity has ever seen." But is this part of a concerning trend? Microsoft plans to absorb its excess heat with a steady flow of air and, as needed, evaporated drinking water. Use of the latter is projected to reach more than 50 million gallons every year. That might be a burden in the best of times. As of 2023, it seemed absurd. Phoenix had just endured its hottest summer ever, with 55 days of temperatures above 110 degrees. The weather strained electrical grids and compounded the effects of the worst drought the region has faced in more than a millennium. The Colorado River, which provides drinking water and hydropower throughout the region, has been dwindling. Farmers have already had to fallow fields, and a community on the eastern outskirts of Phoenix went without tap water for most of the year... [T]here were dozens of other facilities I could visit in the area, including those run by Apple, Amazon, Meta, and, soon, Google. Not too far from California, and with plenty of cheap land, Greater Phoenix is among the fastest-growing hubs in the U.S. for data centers.... Microsoft, the biggest tech firm on the planet, has made ambitious plans to tackle climate change. In 2020, it pledged to be carbon-negative (removing more carbon than it emits each year) and water-positive (replenishing more clean water than it consumes) by the end of the decade. But the company also made an all-encompassing commitment to OpenAI, the most important maker of large-scale AI models. In so doing, it helped kick off a global race to build and deploy one of the world's most resource-intensive digital technologies. Microsoft operates more than 300 data centers around the world, and in 2021 declared itself "on pace to build between 50 and 100 new datacenters each year for the foreseeable future...." Researchers at UC Riverside estimated last year... that global AI demand could cause data centers to suck up 1.1 trillion to 1.7 trillion gallons of freshwater by 2027. A separate study from a university in the Netherlands, this one peer-reviewed, found that AI servers' electricity demand could grow, over the same period, to be on the order of 100 terawatt hours per year, about as much as the entire annual consumption of Argentina or Sweden... [T]ensions over data centers' water use are cropping up not just in Arizona but also in Oregon, Uruguay, and England, among other places in the world. The article points out that Microsoft "is transitioning some data centers, including those in Arizona, to designs that use less or no water, cooling themselves instead with giant fans." And an analysis (commissioned by Microsoft) on the impact of one building said it would use about 56 million gallons of drinking water each year, equivalent to the amount used by 670 families, according to the article. "In other words, a campus of servers pumping out ChatGPT replies from the Arizona desert is not about to make anyone go thirsty."

Read more of this story at Slashdot.

Researchers Create AI Worms That Can Spread From One System to Another

✇Slashdot
Par : EditorDavid
3 mars 2024 à 20:45
Long-time Slashdot reader Greymane shared this article from Wired: [I]n a demonstration of the risks of connected, autonomous AI ecosystems, a group of researchers has created one of what they claim are the first generative AI worms — which can spread from one system to another, potentially stealing data or deploying malware in the process. "It basically means that now you have the ability to conduct or to perform a new kind of cyberattack that hasn't been seen before," says Ben Nassi, a Cornell Tech researcher behind the research. Nassi, along with fellow researchers Stav Cohen and Ron Bitton, created the worm, dubbed Morris II, as a nod to the original Morris computer worm that caused chaos across the Internet in 1988. In a research paper and website shared exclusively with WIRED, the researchers show how the AI worm can attack a generative AI email assistant to steal data from emails and send spam messages — breaking some security protections in ChatGPT and Gemini in the process...in test environments [and not against a publicly available email assistant]... To create the generative AI worm, the researchers turned to a so-called "adversarial self-replicating prompt." This is a prompt that triggers the generative AI model to output, in its response, another prompt, the researchers say. In short, the AI system is told to produce a set of further instructions in its replies... To show how the worm can work, the researchers created an email system that could send and receive messages using generative AI, plugging into ChatGPT, Gemini, and open source LLM, LLaVA. They then found two ways to exploit the system — by using a text-based self-replicating prompt and by embedding a self-replicating prompt within an image file. In one instance, the researchers, acting as attackers, wrote an email including the adversarial text prompt, which "poisons" the database of an email assistant using retrieval-augmented generation (RAG), a way for LLMs to pull in extra data from outside its system. When the email is retrieved by the RAG, in response to a user query, and is sent to GPT-4 or Gemini Pro to create an answer, it "jailbreaks the GenAI service" and ultimately steals data from the emails, Nassi says. "The generated response containing the sensitive user data later infects new hosts when it is used to reply to an email sent to a new client and then stored in the database of the new client," Nassi says. In the second method, the researchers say, an image with a malicious prompt embedded makes the email assistant forward the message on to others. "By encoding the self-replicating prompt into the image, any kind of image containing spam, abuse material, or even propaganda can be forwarded further to new clients after the initial email has been sent," Nassi says. In a video demonstrating the research, the email system can be seen forwarding a message multiple times. The researchers also say they could extract data from emails. "It can be names, it can be telephone numbers, credit card numbers, SSN, anything that is considered confidential," Nassi says. The researchers reported their findings to Google and OpenAI, according to the article, with OpenAI confirming "They appear to have found a way to exploit prompt-injection type vulnerabilities by relying on user input that hasn't been checked or filtered." OpenAI says they're now working to make their systems "more resilient." Google declined to comment on the research.

Read more of this story at Slashdot.

Homeless Man Tries to Steal Waymo Robotaxi in Los Angeles

✇Slashdot
Par : EditorDavid
4 mars 2024 à 08:34
A homeless man "was taken into custody on suspicion of grand theft auto," reports the Los Angeles Times, "after police said he tried to steal a Waymo self-driving car in downtown Los Angeles on Saturday night." The man entered and tried to operate a Waymo vehicle that had stopped to let out a passenger at the corner of 1st and Main at 10:30 p.m., Los Angeles Police Department detective Meghan Aguilar said. After the man, whom a Waymo spokesman described as an "unauthorized pedestrian," entered the vehicle, the company's Rider Support team instructed him to exit the car. When he did not, the company contacted the police, "who were then able to remove and arrest" the man, said Chris Bonelli, a Waymo spokesman... No injuries were reported by the rider, and there was no damage to the vehicle, Bonelli said. The car was stationary during the entire incident because an unauthorized person was identified by the company to be in the vehicle, according to Waymo.

Read more of this story at Slashdot.

India Reverses AI Stance, Requires Government Approval For Model Launches

✇Slashdot
Par : msmash
4 mars 2024 à 10:17
An anonymous reader shares a report: India has waded into global AI debate by issuing an advisory that requires "significant" tech firms to get government permission before launching new models. India's Ministry of Electronics and IT issued the advisory to firms on Friday. The advisory -- not published on public domain but a copy of which TechCrunch has reviewed -- also asks tech firms to ensure that their services or products "do not permit any bias or discrimination or threaten the integrity of the electoral process." Though the ministry admits the advisory is not legally binding, India's IT Deputy Minister Rajeev Chandrasekhar says the notice is "signalling that this is the future of regulation." He adds: "We are doing it as an advisory today asking you to comply with it." In a tweet Monday, Chandrasekhar said the advisory is aimed at "untested AI platforms deploying on the India internet" and doesn't apply to startups. About-face from India's position on AI a year ago.

Read more of this story at Slashdot.

Anthropic Releases New Version of Claude That Beats GPT-4 and Gemini Ultra in Some Benchmark Tests

✇Slashdot
Par : msmash
4 mars 2024 à 16:40
Anthropic, a leading artificial intelligence startup, unveiled its Claude 3 series of AI models today, designed to meet the diverse needs of enterprise customers with a balance of intelligence, speed, and cost efficiency. The lineup includes three models: Opus, Sonnet, and the upcoming Haiku. From a report: The star of the lineup is Opus, which Anthropic claims is more capable than any other openly available AI system on the market, even outperforming leading models from rivals OpenAI and Google. "Opus is capable of the widest range of tasks and performs them exceptionally well," said Anthropic cofounder and CEO Dario Amodei in an interview with VentureBeat. Amodei explained that Opus outperforms top AI models like GPT-4, GPT-3.5 and Gemini Ultra on a wide range of benchmarks. This includes topping the leaderboard on academic benchmarks like GSM-8k for mathematical reasoning and MMLU for expert-level knowledge. "It seems to outperform everyone and get scores that we haven't seen before on some tasks," Amodei said. While companies like Anthropic and Google have not disclosed the full parameters of their leading models, the reported benchmark results from both companies imply Opus either matches or surpasses major alternatives like GPT-4 and Gemini in core capabilities. This, at least on paper, establishes a new high watermark for commercially available conversational AI. Engineered for complex tasks requiring advanced reasoning, Opus stands out in Anthropic's lineup for its superior performance. Sonnet, the mid-range model, offers businesses a more cost-effective solution for routine data analysis and knowledge work, maintaining high performance without the premium price tag of the flagship model. Meanwhile, Haiku is designed to be swift and economical, suited for applications such as consumer-facing chatbots, where responsiveness and cost are crucial factors. Amodei told VentureBeat he expects Haiku to launch publicly in a matter of "weeks, not months."

Read more of this story at Slashdot.

Hyundai booste encore la batterie de son Ioniq 5 au lieu de baisser son prix

✇Numerama.com - Magazine
Par : Raphaëlle Baut
4 mars 2024 à 16:15

C’est déjà la deuxième fois que le constructeur coréen change la batterie de son Ioniq 5. Lancé avec une batterie de 72,6 kWh, le modèle est désormais annoncé avec 84 kWh de capacité. Or, cela ne change pas grand-chose pour ce modèle.

Gartner Predicts Search Engine Volume Will Drop 25% by 2026, Due To AI Chatbots and Other Virtual Agents

✇Slashdot
Par : msmash
4 mars 2024 à 17:20
Gartner: By 2026, traditional search engine volume will drop 25%, with search marketing losing market share to AI chatbots and other virtual agents, according to Gartner. "Organic and paid search are vital channels for tech marketers seeking to reach awareness and demand generation goals," said Alan Antin, Vice President Analyst at Gartner. "Generative AI (GenAI) solutions are becoming substitute answer engines, replacing user queries that previously may have been executed in traditional search engines. This will force companies to rethink their marketing channels strategy as GenAI becomes more embedded across all aspects of the enterprise." With GenAI driving down the cost of producing content, there is an impact around activities including keyword strategy and website domain authority scoring. Search engine algorithms will further value the quality of content to offset the sheer amount of AI-generated content, as content utility and quality still reigns supreme for success in organic search results. There will also be a greater emphasis placed on watermarking and other means to authenticate high-value content. Government regulations across the globe are already holding companies accountable as they begin to require the identification of marketing content assets that AI creates. This will likely play a role in how search engines will display such digital content.

Read more of this story at Slashdot.

Qualcomm Launches First True 'App Store' For AI With 75 Free Models

✇Slashdot
Par : BeauHD
5 mars 2024 à 01:10
Wayne Williams reports via TechRadar: Qualcomm has unveiled its AI Hub, an all-inclusive library of pre-optimized AI models ready for use on devices running on Snapdragon and Qualcomm platforms. These models support a wide range of applications including natural language processing, computer vision, and anomaly detection, and are designed to deliver high performance with minimal power consumption, a critical factor for mobile and edge devices. The AI Hub library currently includes more than 75 popular AI and generative AI models including Whisper, ControlNet, Stable Diffusion, and Baichuan 7B. All models are bundled in various runtimes and are optimized to leverage the Qualcomm AI Engine's hardware acceleration across all cores (NPU, CPU, and GPU). According to Qualcomm, they'll deliver four times faster inferencing times. The AI Hub also handles model translation from the source framework to popular runtimes automatically. It works directly with the Qualcomm AI Engine direct SDK and applies hardware-aware optimizations. Developers can search for models based on their needs, download them, and integrate them into their applications, saving time and resources. The AI Hub also provides tools and resources for developers to customize these models, and they can fine-tune them using the Qualcomm Neural Processing SDK and the AI Model Efficiency Toolkit, both available on the platform.

Read more of this story at Slashdot.

New Topaz Photo AI version 2.4 released

✇Photo Rumors
Par : PR admin
5 mars 2024 à 00:46


Topaz Labs officially released Photo AI version 2.4:

A new workflow

Image processing is the core of our application. We separated the tools and allowed them to run in any order, to put you in control. Now, it’s much easier to see how your image is improved step by step, and we hope this transparency leads to more creative outcomes. You can start editing a photo by adding just the enhancements you want -OR- run autopilot to analyze your image and start with our recommendations. Either way, you can edit, add, and remove enhancements as you go.

Control the processed area for each enhancement.

You decide which parts of the image are processed. The new design allows you to select which areas of the image are processed by each enhancement. Some examples of how you can use the new selection feature is adjusting the color balance of your subject, or selectively denoising only the background. The theme of this release is putting more control in your hands to shape images the way you want.

Goodbye to toggles, hello stacked enhancements.

Now you can use as many enhancements on your image as you want. For example, you can sharpen your subject and background with separate enhancements to be able to control the effect individually. We hope this unlocks new creative freedom and we can’t wait to see what y’all make! There are a few exceptions to this feature but we will continue to work on enabling the flexibility of stacking enhancements in future releases.

Customizable interface

We’re just taking our first steps toward customization in this release. You can now drag the enhancement panels out and place them where they work best for you. The team is working hard on more interface organization features and you can expect to see greater flexibility in future releases.

We’ve revamped how filters are processed to make Photo AI more responsive.

Creating a flexible processing order allowed us to dramatically improve the efficiency of processing:

  • Preview updates faster when new edits are added
  • Switching between your current edit and previous edits is much faster
  • Background processing is enabled for computers with 24GB or more RAM when preview is updated (processor is not idle)

All this means editing your image feels faster compared to previous versions.

Crop it like it’s hot!

We know it’s been a long time coming. We’ve overhauled our crop tool to make it more responsive and added critical features like straightening. We’ve also fixed quite a few bugs. The new crop tool loads and saves much faster, and no longer runs Autopilot after for a snappy experience.

The Remove tool can now be used like other filters.

In the new workflow, you can use Remove at any point. It will lock previous edits, which is necessary to maintain consistent output from our generative AI. But other edits can be added after so it does not disrupt your image processing. We updated the controls so they are easier to use. Remove can also be easily edited or deleted from the stack.

What’s next

We are committed to adding new features that will reshape your favorite enhancing tool. Some of the most important work we’re doing:

  • Adding new tools and enhancements
  • Enabling more flexibility in how enhancements work
  • New customization features
  • Improving processing performance

The post New Topaz Photo AI version 2.4 released appeared first on Photo Rumors.

Copilot Pane As Annoying As Clippy May Pop Up In Windows 11

✇Slashdot
Par : BeauHD
5 mars 2024 à 23:20
Richard Speed reports via The Register: Copilot in Windows is set to get even more assertive after Microsoft added a function that makes the AI assistant's window pop up after a user's cursor hovers over the icon in the task bar. [...] Windows Insiders on the Beta Channel â" with the option to get the latest updates turned on â" will soon find themselves on the receiving end of what Microsoft calls "a new hover experience for Copilot in Windows" from build 22635.3276. If your mouse cursor happens to drift over to the Copilot icon on the taskbar, the Copilot pane will open to make users aware of the delights on offer. The result, we suspect, will be to educate users in the art of switching off the function. Much like Widgets, which will also make its unwanted presence felt should a user move a mouse over its icon. A swift hop into taskbar settings is all it takes to make the icons disappear, for now at least. The new feature is being piloted but considering the proximity of the Beta Channel to Release Preview, there is every chance the pop-up will, er, pop up in a release version of Windows before long.

Read more of this story at Slashdot.

Attaqué par Elon Musk, OpenAI riposte en dévoilant ses mails privés

✇Numerama.com - Magazine
Par : Julien Lausson
6 mars 2024 à 11:49

elon musk

Les tensions entre OpenAI et Elon Musk montent d'un cran. Après la plainte du milliardaire américain, les fondateurs de l'entreprise spécialisée en IA dévoilent des mails de leur ancien partenaire. Ils entendent souligner le double discours de l'entrepreneur.

JPMorgan's AI-Aided Cashflow Model Can Cut Manual Work by 90%

✇Slashdot
Par : msmash
6 mars 2024 à 21:40
JPMorgan helped some of its corporate customers slash manual work by almost 90% (alternative source) with its cashflow management tool that runs on AI, bringing the largest US bank one step closer to charging for this service. From a report: "We are going to keep investing into this solution because we see that we're starting to really crack this workflow," said Tony Wimmer, head of data and analytics at JPMorgan's wholesale payments unit, in an interview. Since launching about a year ago, his firm now has about 2,500 clients using the product, he said. The tool, which allows corporate treasuries to analyse and forecast cash flows, has seen "tremendous" interest from its clients who currently use it for free, Wimmer said. His firm is considering charging its customers in the future to use the solution, dubbed Cash Flow Intelligence. The world's biggest banks have been stepping up their use of artificial intelligence with the aim of lifting productivity and reducing costs. JPMorgan's Chief Executive Officer Jamie Dimon has said the technology could eventually allow employers to shrink the workweek to just 3.5 days. JPMorgan set a target of $1 billion in "business value" generated by AI in 2023, and the firm increased that goal to $1.5 billion at its investor day in May.

Read more of this story at Slashdot.

Public Trust In AI Is Sinking Across the Board

✇Slashdot
Par : BeauHD
6 mars 2024 à 23:40
Trust in AI technology and the companies that develop it is dropping, in both the U.S. and around the world, according to new data from Edelman shared first with Axios. Axios reports: Globally, trust in AI companies has dropped to 53%, down from 61% five years ago. In the U.S., trust has dropped 15 percentage points (from 50% to 35%) over the same period. Trust in AI is low across political lines. Democrats trust in AI companies is 38%, independents are at 25% and Republicans at 24%. Tech is losing its lead as the most trusted sector. Eight years ago, technology was the leading industry in trust in 90% of the countries Edelman studies. Today, it is the most trusted in only half of countries. People in developing countries are more likely to embrace AI than those in developed ones. Respondents in France, Canada, Ireland, UK, U.S., Germany, Australia, the Netherlands and Sweden reject the growing use of AI by a three-to-one margin, Edelman said. By contrast, acceptance outpaces resistance by a wide margin in developing markets such as Saudi Arabia, India, China, Kenya, Nigeria and Thailand. "When it comes to AI regulation, the public's response is pretty clear: 'What regulation?'," said Edelman global technology chair Justin Westcott. "There's a clear and urgent call for regulators to meet the public's expectations head on."

Read more of this story at Slashdot.

'AI Prompt Engineering Is Dead'

✇Slashdot
Par : msmash
7 mars 2024 à 15:20
The hype around AI language models has companies scrambling to hire prompt engineers to improve their AI queries and create new products. But new research hints that the AI may be better at prompt engineering than humans, indicating many of these jobs could be short-lived as the technology evolves and automates the role. IEEE Spectrum: Battle and Gollapudi decided to systematically test [PDF] how different prompt engineering strategies impact an LLM's ability to solve grade school math questions. They tested three different open source language models with 60 different prompt combinations each. What they found was a surprising lack of consistency. Even chain-of-thought prompting sometimes helped and other times hurt performance. "The only real trend may be no trend," they write. "What's best for any given model, dataset, and prompting strategy is likely to be specific to the particular combination at hand." There is an alternative to the trial-and-error style prompt engineering that yielded such inconsistent results: Ask the language model to devise its own optimal prompt. Recently, new tools have been developed to automate this process. Given a few examples and a quantitative success metric, these tools will iteratively find the optimal phrase to feed into the LLM. Battle and his collaborators found that in almost every case, this automatically generated prompt did better than the best prompt found through trial-and-error. And, the process was much faster, a couple of hours rather than several days of searching.

Read more of this story at Slashdot.

EA Says Generative AI Could Make It 30% More Efficient

✇Slashdot
Par : msmash
7 mars 2024 à 16:40
EA CEO Andrew Wilson believes generative AI will "revolutionize" the gaming industry over the next five years. He predicts that the technology will allow for more efficient content creation, reducing development time from months to days. From a report: Greater efficiency coupled with "deeper, more immersive experiences" will lead to significant audience expansion over the next few years and provide a "multi-billion dollar" growth opportunity, he said. Wilson said that in the past it might take six months to build an in-game sports stadium. Over the last 12 months, that time has shrunk to six weeks, and over the coming years it could maybe be cut to six days. And while FIFA 23 has 12 run cycles for how the players move in the game, EA Sports FC 24 has 1,200 created with generative AI. Over the next five years, Wilson hopes that generative AI will make EA's development 30% more efficient, help grow its 700 million-strong player base by "at least" 50%, and lead to players spending 10-20% more money on its games. "What we've seen every time there's been a meaningful technological advancement in media and in technology, where you are able to democratise an industry and hand it over to the population at large, incredible things happen," he said.

Read more of this story at Slashdot.

❌
❌