GitHub Spark est un outil alimenté par l’IA qui permet de créer et de partager des micro-applications personnalisées, appelées “sparks”, sans avoir besoin d’écrire ou de déployer du code. Ces applications peuvent être utilisées directement sur vos appareils de bureau et mobiles. GitHub Spark offre un éditeur basé sur le langage naturel pour décrire et affiner vos idées, un environnement d’exécution géré pour héberger vos sparks avec accès au stockage de données, aux thèmes et aux modèles de langage, ainsi qu’un tableau de bord compatible PWA pour gérer et lancer vos sparks depuis n’importe où. De plus, vous pouvez partager vos sparks avec d’autres utilisateurs, en contrôlant les autorisations de lecture ou d’écriture, et permettre à chacun de les adapter à ses préférences.
How did supermassive black holes end up at the center of every galaxy? A while back, it wasn't that hard to explain: That's where the highest concentration of matter is, and the black holes had billions of years to feed on it. But as we've looked ever deeper into the Universe's history, we keep finding supermassive black holes, which shortens the timeline for their formation. Rather than making a leisurely meal of nearby matter, these black holes have gorged themselves in a feeding frenzy.
With the advent of the Webb Space Telescope, the problem has pushed up against theoretical limits. The matter falling into a black hole generates radiation, with faster feeding meaning more radiation. And that radiation can drive off nearby matter, choking off the black hole's food supply. That sets a limit on how fast black holes can grow unless matter is somehow fed directly into them. The Webb was used to identify early supermassive black holes that needed to have been pushing against the limit for their entire existence.
But the Webb may have just identified a solution to the dilemma as well. It has spotted a black hole that appears to have been feeding at 40 times the theoretical limit for millions of years, allowing growth at a pace sufficient to build a supermassive black hole.
Ubisoft traverse une période complexe marquée par d’importantes pressions financières et sociales. Face à des résultats décevants et à des tensions internes, l’entreprise cherche à maintenir un équilibre délicat entre rentabilité et innovation, dans un secteur où la créativité est fondamentale. Des résultats financiers en déclin et un plan de réduction des coûts Pour le …
On October 17th, 2024, Jason McDowell (The AR Show), Jeri Ellsworth (Tilt Five), David Bonelli (Pulsar), Bradley Lynch (SadlyItsBradley), and I recorded a 2-hour roundtable discussion about the recent announcements of the Snap Spectacles 5 and Meta Orion optical AR/MR glasses. Along the way, we discussed various related subjects, including some about the Apple Vision Pro.
I divided the two-hour discussion into three parts, with this being the last and, I think, in some ways, the most interesting part. We discussed more of the motivations for Meta and Snap announcing these concept products. Brad shared some information he has heard about issues with Meta’s EMG wristband control, and Jeri shared some of her past experiences with EMG-type control.
Warning: Much of this video is based on opinions based on prior tech business experience and second-hand information.
Part 3 of the AR Roundtable Video
As per my prior companion articles for this roundtable video, I will try to fill in some details from our on-the-fly discussions. Below is a link to Part 3 of the roundtable, followed by a brief description of each section.
0:00 Introduction of the panelist (repeated from part 1)
Jason asked Brad about what he thought about the wide FOV on Meta’s Orion. However, Brad’s answer was more directed at what he thought about Meta’s interest in VR and Meta’s motivation in AR. Brad said that a small screen/FOV will forever be seen as an accessory to a smartphone and that Meta wants a big screen/FOV because it fits better with Meta’s long-term play to try and replace the smartphone.
Brad’s comment led the various panel members to discuss Meta’s motivations in AR and the concept of what Jeri called a “Walled Gardens” (to have an ecosystem that one company controls and can defend against competitors). I talked about “intersection points,” which are times when technology has advanced to the point where a set of technologies has progressed to the point that can enable a new category of products such as PCs, Laptops, and Smartphones.
I brought up the question of whether AI is an intersection point, as many people think today. The next question is whether someone can set up a “walled garden” or whether multiple companies can have their own AI, where it becomes more of a commodity that many companies have rather than a defensible barrier to entry into a product category. Can someone “spin up” the “buyers bring sellers and sellers bring buyers” barriers as we saw with eBay and Google search? Or the way that Amazon’s shipping network is one of their advantages and a barrier to other entering their space.
David talks about “content-driven” versus “user-driven” interaction. With content-driven interaction, you can live with a smaller FOV, but with “user-driven,” you want a wider FOV and eye tracking. With a user-driven approach, information is provided based on what the user is looking at, and this fits Meta’s revenue models better. Jeri comments that the wider FOV will help Meta grab the user’s attention.
Brad has been following EMG wristbands for some time and has some interesting feedback from people who have used Meta’s EMG wristband. According to Brad’s sources, Meta’s wristband is far from being ready as a product. First, it can have problems working with people who have a lot of fat or hair around their wrists. He said that for some people, the wristband has to be uncomfortably tight if it works at all. The next issue is reliability in detecting gestures. Brad’s sources said that Meta had to greatly simplify the set of gestures and make them more overt to improve recognition. Like most gesture-based methods, there is going to be a problem of recovering from false positives.
Jeri shared some of her experiences working with EMG devices, particularly problems with lag and how they make user control difficult. She also discussed how the reliability problems as sensors shift/move with use and if the user sweats. Brad added that the detection is very sensitive to electrical interference; even a smartwatch on top of the wristband can interfere with its operation, and he amplified the issue of latency.
I discussed that any form of gesture-based input ties up the hands, which is a problem for many applications that need to be “hands-free.”
17:56 Tethering to a Compute Pack and the need for a second device
I generally agree that “wireless tethering” is infinitely better than a wired cable back to a processing pod or phone. Many companies, including Meta, with Snap Spectacles 5 being the recent exception, move weight from the headset to a “dongle” battery and/or compute pod.
This discussion brought up Meta’s overall strategy and the issue that Apple would seem to have an “ace-card” with their “wireless dongle” being an iPhone people already own. How could Meta get someone to buy their AR glasses with the compute pod and forgo their smartphone?
18:52 Replacing the Smartphone and Business Models (Meta, Apple, and Snap)
This discussion brought up Meta’s overall strategy and the issue that Apple would seem to have an “ace-card” with their “wireless dongle” being an iPhone people already own. How could Meta get someone to buy their AR glasses with the compute pod and forgo their smartphone? Similar questions were raised about Snap Spectacles. Jason questions whether Meta has to develop a smartphone of its own, which has big barriers to entry.
We then go on with a far-ranging discussion of various business and market issues and motivations.
37:28 Eye Glow and Captured Light with Diffractive Waveguides
David discussed a bit about the social issues about (far field) light either being captured and reflective and eye glow from the waveguides. Jeri suggested that companies might have to try to own the issue and try to turn it from a negative to a positive.
39:42 Orion’s Eye Glow vs. Apple Vision Pro’s Eyesight
Brad commented that people’s eyes look a lot like Apple Vision Pro’s EyeSight feature when wearing Orion due to the eye glow.
David points out that AR glasses are not “portable” if you have to remove and store them when not in use and that people are not going to wear big and bulky glasses all day.
The price of the largest cryptocurrency is flirting with all-time highs ahead of an election that may be a turning point for the future of digital assets. Former President Donald Trump said he would take crypto “to the moon” at a convention in July, and cement the US as a global leader. Vice President Kamala Harris’ campaign also warmed up to the industry in recent weeks, saying her administration would foster innovation. In case you couldn’t guess, “Big Crypto” has reportedly become the largest lobbyist this election cycle.
“Looking at the election, both campaigns have hinted at pro-crypto policies,” said Blake Morgan, managing partner at the investment firm Mineral Vault. Bitcoin rallied 13% in October alone. “A more crypto-friendly environment would likely have a positive impact.”
Lunar Landings
All of that excitement has led investors to pour about $2.22 billion into exchange-traded funds that invest in underlying bitcoin last week. BlackRock took the lion’s share, with more than $2.15 billion landing in the iShares Bitcoin Trust ETF (IBIT). Even more impressive: $872 million came on Wednesday alone, shattering previous all-time highs for single-day inflows, according to Bloomberg.
“$IBIT took in more cash than any other ETF in the world over the past week,” Bloomberg Intelligence senior analyst Eric Balchunas said in a post on X. It beat the more than 13,000 funds in the universe, including powerhouses like the Vanguard S&P 500 ETF and the iShares Core S&P 500 ETF, he added.
Keep in mind, IBIT is just 9 months old. “[It’s] a huge moment for the industry,” Morgan said.
What’s FOMO? It’s not all rainbows and rocket ships, however. Given crypto’s perceived risks, it’s been a hard sell for financial advisors that are much more comfortable playing it safe with client cash. Confidence has yet to fully recover from the 2022 crash, and investments into the sector are still playing catch-up with pre-“crypto winter” levels. VC funding into crypto firms continues to decline; this year has seen the smallest average investment sizes since 2017, according to a report by Galaxy.
“Bitcoin remains volatile, and the swings can be intense,” Morgan told The Daily Upside. Still, “bitcoin has shown unique growth potential that’s hard to match in traditional markets.”
Les noyaux atomiques, composés de protons et de neutrons, cachent en leur cœur des quarks et des gluons. Ces derniers, encore difficiles à étudier, ont longtemps échappé aux scientifiques.
Au...
If you’re into pushing tech boundaries from home, this one’s for you. Redditor [mi_kotalik] has crafted ‘Zero’, a custom pair of DIY augmented reality (AR) glasses using a Raspberry Pi Zero. Designed as an affordable, self-contained device for displaying simple AR functions, Zero allows him to experiment without breaking the bank. With features like video playback, Bluetooth audio, a teleprompter, and an image viewer, Zero is a testament to what can be done with determination and creativity on a budget. The original Reddit thread includes videos, a build log, and links to documentation on X, giving you an in-depth look into [mi_kotalik]’s journey. Take a sneak peek through the lens here.
Creating Zero wasn’t simple. From designing the frame in Tinkercad to experimenting with transparent PETG to print lenses (ultimately switching to resin-cast lenses), [mi_kotalik] faced plenty of challenges. By customizing SPI displays and optimizing them to 60 FPS, he achieved an impressive level of real-time responsiveness, allowing him to explore AR interactions like never before. While the Raspberry Pi Zero’s power is limited, [mi_kotalik] is already planning a V2 with a Compute Module 4 to enable 3D rendering, GPS, and spatial tracking.
Des grains d'astéroïde seraient-ils à l'origine des premières briques de la vie sur Terre ? C'est ce que suggèrent les récentes découvertes faites à partir de poussières venues de l'espace....
Entre communautés ciblées et les canaux privilégiés, Launchmetrics dresse le portrait des stratégies des entreprises lors de cette saison de consommation à ne pas rater, du Black Friday au Cyber Monday en passant par les fêtes de fin d'année.
In recent years and with the many advances in technology, online casinos have been increasingly focusing on providing a more personalised gaming experience to players.
As the competition between online casinos gets tougher, casinos are seeking new and innovative ways to create personalised offers, rewards and bonuses that fit the individual’s preferences, creating a fully customised environment for each player.
This has led to the rise of dynamic bonuses, where machine learning algorithms are designed to predict player preferences and provide them with relevant offers that hit their target. From a deep dive into the various casino bonus structures and player experiences, machine learning is proving itself to be a powerful tool in shaping the future of player retention and satisfaction.
The main role of machine learning in dynamic bonuses
Machine learning allows casinos to quickly and promptly analyse huge amounts of data on player behaviour, from game choices and spending habits to the lengths of gaming sessions and win-or-loss ratios.
By using these very important data points, machine learning models can quickly identify patterns and also predict what types of casino bonuses or rewards will appeal the most to specific players. For example, a player who enjoys playing online slot games is more likely to appreciate a free spins bonus, while a player who prefers RNG card and table games or live dealer games might prefer a match deposit bonus offer.
‘Machine learning allows casinos to delve into a player’s journey and create tailormade bonuses in real-time, which can make a huge difference when it comes to engagement and satisfaction,’ says Jonas Kyllönen, Mr. Gamble’s co-founder and casino expert.
By better aligning with these predictive insights, casinos can not only optimise their marketing efforts but also ensure they provide a value-driven experience for their players.
Casino bonuses: traditional vs dynamic approach
Traditional casino bonuses are often created with a ‘one-size-fits-all’ mentality, where the same offers are available to all players. This can lead to the casinos missing engagement opportunities since these offers do not take into consideration individual preferences.
Dynamic bonuses, on the other hand, take into account data-driven insights, offering more personalised rewards that are more likely to appeal to each player. This results in a better engagement strategy that not only attracts players but is likely to see them return for more.
By using ML_driven dynamic bonuses, online casino sites can attract players in unique and very customised ways. Highlighting popular £10 deposit bonuses in the UK, for example, can attract new players looking for low-commitment entry points, while experienced players might benefit from loyalty rewards of exclusive promotions.
How machine learning models in dynamic bonuses work
Machine learning models designed to predict player preferences are typically built on three core components: data collection, pattern recognition and personalised engagement. You can find out more about each one below:
- Data collection: Casinos collect data across all player interactions, including games played, frequency, the device used and the time of day for such gaming sessions.
- Pattern recognition: By analysing this data, ML algorithms detect various behavioural patterns, recognising tendencies that might predict potential future preferences.
- Personalised engagement: With patterns identified, the model can also recommend specific bonuses or promotions to individual players, increasing the likelihood of them accepting the offer and engaging.
These data-driven insights are designed to address player preferences, from offering tailor-made welcome bonuses to designing loyalty programs that align perfectly with the player’s gaming history.
A player that makes small deposits on a frequent basis might be more attracted to the best PayPal casinos in the UK, where they can benefit from safe and smooth transactions, along with snapping up casino bonuses perfectly designed for small deposits.
How online casinos use dynamic bonuses
The power of dynamic bonuses lies in their ability to match casino bonus offers with individual player profiles.
Some online casinos are now using machine learning to send real-time bonus offers based on the player’s current activity. If the player is on a winning streak, they might receive an extra casino bonus for example in a bid to encourage them to try a different game category.
On the other hand, a player who is on a losing streak might be offered cashback or a low-risk casino bonus to maintain engagement levels.
Another example entails highlighting the most attractive casino bonuses based on deposit methods. For example, casinos can create specific casino bonuses towards players opting for PayPal. Offering unique bonuses for PayPal players is not only about luring them to the site but also reassuring them with fast and secure transactions.
Here are some of the most popular UK dynamic bonuses created by machine learning:
The future of dynamic bonuses in the casino industry
As machine learning continues to develop and also advance, the future of dynamic bonuses promises to become even more sophisticated and intricate. Thanks to the integration of artificial intelligence and machine learning algorithms thrown into the mix, online casinos will be able to anticipate player needs with greater precision.
In such a competitive and rife casino environment, a new casino must do its utmost to stand out in a competitive market. One way of doing this is via personalised features such as dynamic bonuses.
Picture yourself logging into your favourite online casino and receiving a casino bonus offer tailored to your favourite game type, time of day or even previous win/loss partners: These sorts of personalised experiences are becoming a standard for many players who crave that unique and value-driven interaction with online casino sites.
Not only do such experiences foster engagement, but they also encourage a sense of loyalty.
As an innovator in 3D scanning technology, Revopoint has unveiled its latest product, the Revopoint MetroX 3D scanner. The Kickstarter campaign started on October 15th and will run until November 14th, offering MetroX at only $689 (MSRP $999) now.
Four Scanning Modes
MetroX is a versatile 3D scanner that combines blue laser and blue structured light technology. Its four scanning modes give users the right tools to optimally capture every aspect of small to medium objects, from flat areas to complex details.
Cross Lines laser: Fourteen crossed lines rapidly scan larger surfaces at speeds of up to 800,000 points/second. No scanning spray is needed for glossy or black areas.
Parallel Lines laser: Seven high-powered parallel blue laser lines are projected to accurately capture fine details, edges, and complex surfaces.
Full-field structured light: Using blue structured light, it offers high-speed scanning at up to 7 million points/second for efficient point cloud capture. This mode also enables marker-free scanning of feature-rich objects, simplifying the process and reducing setup time.
Auto Turntable: The hands-free setup lets users place MetroX on a desk and automatically scan objects using the Dual-axis Turntable and full-field structured light. This mode ensures reliable point cloud and color capture every time.
High Precision and Accuracy
MetroX captures feature-rich and featureless objects with remarkable fidelity—up to 0.01 mm precision, 0.03 mm accuracy, and a fused point distance of up to 0.05 mm.
Additionally, the included high-end lithography calibration plate features pattern accuracy of up to 10 μm, making it easy to calibrate the MetroX 3D scanner and maintain reliable accuracy.
Seamless Workflow
The free companion software, Revo Scan, streamlines the scanning and post-processing workflow, covering point cloud fusion, meshing, editing, and import/export options. It can also offer a wide variety of file formats for full compatibility with third-party software like Quicksurface, Geomagic Wrap, and Geomagic Control.
Professional Solutions
MetroX offers professionals an enhanced reverse engineering, quality control, additive manufacturing, and rapid prototyping 3D modeling experience with metrology-grade accuracy at an appealing price.
What’s in the Box?
The Revopoint MetroX 3D scanner comes in a smart-looking aluminum carry case that houses the scanner and all its accessories, including a Dual-axis Turntable, a tripod, markers, a calibration board, a power adapter, a USB Type-A to Type-C cable, a scanner cradle and other essential tools. Additionally, a quick start guide is included to ensure that users have everything necessary to begin using the scanner promptly.
The Advanced Edition of the MetroX 3D scanner comes with the Marker Block Kit, which includes 10 magnetic dome markers, 4 magnetic pyramid markers, 10 magnetic ball markers, 4 magnetic bar markers, and additional 500 markers, all specifically designed to enhance and simplify your scanning process.
Le truc marrant avec l’AI Pin, c’est son tarif. 699$ HT et l’obligation de prendre un abonnement cellulaire dédié à 24$ HT. Ce qui nous fait une première année à 990$ HT soit quelque chose comme 1100€. Pour ce prix, on peut s’attendre à un appareil de qualité… mais en fait c’est pas grand chose d’autre qu’un outil qui sert de pont vers un serveur pour lancer des questions et vous communiquer ensuite les réponses obtenues. Le fait qu’il soit incapable de faire cette tâche de manière efficace en étant mauvais dans la capture de son environnement et fort lent dans ses réponses est déjà en soi un problème. Le fait que l’appareil ne soit pas bien pensé, chauffe et pendouille sur vos vêtements est un autre problème. Mais on apprend que Humane vient de proposer un rappel de son chargeur pour un défaut de conception.
Tous les clients de la marque qui ont acheté cette option (oui c’est une option à 129$HT en plus du prix de base) ne doivent pas l’utiliser car la batterie intégrée pourrait surchauffer et détruire l’appareil que l’objet est censé recharger. Plus de 10 000 chargeurs sont donc rappelés par la marque. Une fois le rappel validé l’utilisateur , je ne sais pas si ce mot est valable, sera remboursé ou pourra espérer un chargeur de remplacement dans les 3 à 6 mois…
C’est en réalité une bonne nouvelle car j’imagine que plus personne n’ose trop porter ce bidule à son torse de peur de passer pour le dernier des crétins. Du coup il y a une petite chance de pouvoir récupérer quelques dollars dans ce cauchemar technologique.
Regardez autour de vous, on trouve des dizaines et des dizaines de matériels qui emploient des chargeurs de ce type. Des casques audio, des micro cravates, des appareils en tous genres qui ont adopté des chargeurs mobiles. Je ne parlerais pas des batteries externes USB car là ce sont des centaines de marques et de modèles qui existent sur le marché. Apple propose son chargeur d’AirPod Pro à 109€ TTC, son chargeur classique est à 89€ TTC. Un étui de charge DJI Mic 2 est commercialisé à 69€ TTC.
On est loin des 129$ HT (140€ TTC) du chargeur d’Humane. Et encore, c’est le prix quand on achète l’appareil et le chargeur ensemble, sinon c’est vendu 149$ HT (164€ TTC). Comme peut-on construire un chargeur que l’on va vendre à 160€ et ne pas s’assurer qu’il ne présentera aucun risque pour son produit vendu 699$ HT et que l’on présente comme le fleuron de l’intelligence humaine ? Cela me dépasse. Des marques absolument « Noname » arrivent à fournir des chargeurs de ce type sans soucis. Des marques internationales proposent des produits de qualité pour moins cher et Humane arrive à se rater sur cet accessoire.
La leçon à tirer de tout cela est toujours la même. Il ne faut jamais se précipiter sur des produits qui surfent sur la dernière mode, c’est très rarement signe d’un produit parfaitement fiable. Ni d’un point de vue logiciel, ni d’un point de vue matériel.
Many PCB vendors now provide flexible PCB manufacturing services, but Murata goes further with stretchable PCB technology that’s not only bendable but can be twisted and stretched to better fit on the body for bio-sensing and medical applications even on parts such as an elbow.
Traditional bio-monitoring sensors have some limitations. For example, they can become unstuck when the body moves, damage the delicate skin of infants and the elderly, data may be distorted due to body movement, and Murata explains “there is a risk of a decline in insulation and the occurrence of ion migration when using a thermoplastic polyurethane elastomer (TPU), known as a stretchable base material, in a high humidity environment”. The company’s stretchable printed circuits ((SPC) are supposed to solve or at least mitigate all those issues.
The technology is under development, but Murata still shared some highlights of the technology:
Stretchable electrode printing in compliance with ANSI/AAMI EC12 (in-house test). EC12 is a standard established by the American National Standards Institute (ANSI) and the U.S. Association for the Advancement of Medical Instrumentation (AAMI) that stipulates the specifications required in disposable electrodes used when conducting electrocardiograms and other tests.
About 100 µm per layer
PCB tested with stretchable silver and silver chloride electrodes
Suppresses the occurrence of ion migration, a phenomenon that causes short circuits in metal wiring, pattern wiring, and other elements formed on resin materials due to the application of a constant voltage in a high-humidity environment.
Coating with a safe base material verified with a cytotoxicity test (in-house test based on ISO 10993)
Some applications include stretchable bioelectrode array sheets for bio-potential and bioimpedance measurements such as electrocardiograms (ECGs), electroencephalograms (EEGs), and electromyographs (EMGs), active electrode sheets equipped with an amplifier near the bioelectrode, wireless stretchable wearable patches equipped with BLEs, MCUs, and coin batteries, multi-parameter measurement sheets equipped with various sensors such as thermistors and accelerometers, and shielded soft signal distribution cables.
Stretchable bioelectrode array sheets for bio-potential and bioimpedance measurements
Murata did not provide a timeline for the stretchable printed circuit services, but the company shares that they will be able to design and manufacture interface modules between the human body and medical devices, jointly develop custom products, and provide local support. But they can NOT help customers design medical devices, implement final productization and conduct inspections, conduct clinical trials, and obtain approval as medical devices. More details can be found on Murata’s website and the video below is worth a watch as stretchable PCBs can be twisted and stretched by more than I initially thought.
3D printing has allowed the hobbyist to turn out all sorts of interesting chess sets with either intricate details or things that are too specialized to warrant a full scale injection molded production run. Now, the magic of 3D printing has allowed [Works By Design] to change the game by making pawns that can automatically transform themselves into queens.
Inspired by a CGI transforming chess piece designed by [Polyfjord], [Works By Design] wanted to make a pawn that could transform itself exist in the real world. What started as a chonky setup with multiple springs and a manually-actuated mechanism eventually was whittled down to a single spring, some pins, and four magnets as vitamins for the 3D printed piece.
We always love getting a peek into the trial-and-error process of a project, especially for something with such a slick-looking final product. Paired with a special chess board with steel in the ends, the magnets in the base activate the transformation sequence when they reach the opposite end.
Les liens entre la santé cardiaque et cognitive sont de plus en plus documentés, et une nouvelle étude de l'American Heart Association (AHA), publiée dans Stroke, en apporte une nouvelle preuve....
There are plenty of AI gadgets, but few are truly useful. The Plaud NotePin is an exception and it's proven to be the AI gadget that's right for me. Here's why.
Is ChatGPT a little close-minded, a little old-fashioned in its thinking?
Ask Chanel CEO Leena Nair, who was left disappointed — if not insulted — after she got to experiment with the OpenAI chatbot during her visit to Microsoft's headquarters in Seattle.
Nair, who is the second female global CEO in the iconic fashion house's nearly 115-year history, recalls how she gave ChatGPT a fairly softball image prompt — only for it to screw up badly.
"We're like, 'Show us a picture of a senior leadership team from Chanel visiting Microsoft,'" she said in an interview at the Stanford Graduate School of Business, as quoted by Fortune.
The result? It was "all men in suits," she recalled.
That'd be a mistake at any big corporation, but it's a particularly egregious one to make for Chanel, a brand that mostly caters to women and was founded by one: Coco Chanel, one of the most famous fashion designers in history. You'd think ChatGPT would've scraped a Wikipedia article on that.
"This is Chanel," Nair asserted in the interview. "76 percent of my organization is women. 96 percent of my clients are women. Female CEO," she added, pointing to herself.
"It was a 100 percent male team, not even in fashionable clothes," she roasted. "Like, come on. This is what you've got to offer?"
Some of the half-baked attempts to put guardrails against this have only heightened the obviousness of their shortcomings.
The problem is that AI models are mostly a product of what they're trained on — the internet, replete with repulsive content — and therefore reinforce the same biases that humans hold. When the tech bears the image of being some impartial, superintelligent arbiter of reality, being presented with its overtly sexualized depictions of women, for example, can be whiplash-inducing.
That being said, Nair is a believer in AI, and says that Chanel has been working on getting "AI-ready." But she's ardent that the tech needs to be designed with "ethics and integrity" in mind.
"I constantly talk to my friends in tech, all the CEOs, saying, 'Come on, guys, you gotta make sure that you're integrating a humanistic way of thinking in AI,'" Nair said, per Fortune.
For now, though, we'd suggest that "AI industry" and "ethics" is a bit of an oxymoron.
Scientists are working on turning back the clock on human hearts, allowing patients to grow new heart muscle even after cardiac injury.
As the Wall Street Journal reports, these treatments could represent new hope for patients with heart disease, even after suffering a heart attack.
However, scientists still have a lot to prove, and clinical trials of such treatments in humans are still many years out. Whether any of them turn out into viable ways to treat heart disease, the leading cause of death in the US, remains to be seen.
However, early animal trials are proving to be promising, offering a glimpse of a future in which we could undo heart damage that's currently seen as irreversible.
One approach involves using microRNA to force cardiac cells to start to multiply again. Early experiments by researchers at King's College London involving pigs showed that such a therapy could improve the heart's pumping function, but researchers are still investigating a more effective delivery method.
Other scientists at Scripps Research are using drugs that target proteins that are responsible for cell growth, with the goal of inducing patients' hearts to physically grow larger, as the WSJ notes. Early trials involving mice and pics saw the animals' heart-pumping capacity almost go back to normal after suffering a heart attack.
Some researchers are also working on stem cell therapies that can grow new heart cells, which can then be grafted onto a damaged heart, encouraging it to "remuscularize." University of Southern California stem cell researcher Chuck Murry and his colleagues recently found that their stem cell treatment allowed macaques' hearts to regain full pumping function.
But as the WSJ reports, Murry and his team found that the new cells caused arrhythmias because they were beating to their own rhythms.
Nonetheless, he's hoping to kick off human clinical trials as soon as 2026.
"Society has become OK with the notion of dying from heart disease," he told the WSJ. "It does not have to be this way, because we can do something about this now."
A more experimental approach involves growing entire "personalized" hearts in a lab using cultured human stem cells. Doris Taylor, the CEO of a biotech company called Organamet Bio, told the newspaper that she's hoping to kick off human clinical trials in about five years.
It's possible that none of these treatments will pan out. But it's a promising new avenue of research — and one that just might make a dent in the leading cause of hospitalizations in the US in the not-so-distant future.
Volumetric displays are simply cool. Throw some LEDs together, take advantage of persistence of vision, and you’ve really got something. [Nick Electronics] shows us how its done with his neat little volumetric lamp build.
The concept is simple. [Nick] built a little device to spin a little rectangular array of LEDs. A small motor in the base provides the requisite rotational motion at a speed of roughly 6000 rpm. To get power to the LEDs while they’re spinning, the build relies on wire coils for power transmission, instead of the more traditional technique of using slip rings.
The build doesn’t do anything particularly fancy—it just turns on the whole LED array and spins it. That’s why it’s a lamp, rather than any sort of special volumetric display. Still, the visual effect is nice. We’ve seen some other highly capable volumetric displays before, though. Video after the break.
The company is seeking to patent a system for “multi-task automatic speech recognition.” OpenAI’s patent details a voice-activated AI model that’s able to handle tasks in several different languages.
OpenAI’s tech uses a transformer model, which learns context and relationships between data, outfitted with an encoder and decoder to process streams of audio and turn them into text. The decoder is configured to pick up a “language token,” which specifies the target language for translation, as well as a “task token,” which determines the task that the audio stream is asking for.
Additionally, the transformer model is configured to understand “special-purpose tokens,” which guide it to complete specific tasks, and “timestamp tokens,” which time-align audio to text.
“There are many different tasks that can be performed on the same input audio signal: transcription, such as translation, voice activity detection, time alignment, and language identification,” OpenAI said in the filing.
This system helps the model understand relationships between audio snippets and their corresponding text to make it more efficient at translation and task performance, and uses these specific tokens to hone the model’s skills in specific contexts.
Voice-operated AI has become a priority for OpenAI. The company unveiled its advanced voice mode back in May with the announcement of GPT-4o, and released the feature to an invite-only group in July before opening it up to a wider audience at the end of September.
The model surpasses its standard voice mode, capable of handling interruptions and interpreting emotion in a user’s voice. Additionally, the company unveiled several new tools earlier this month that are capable of fast-tracking voice assistants’ development using only a single set of instructions.
However, the company has run into problems with some of its speech recognition tech. OpenAI’s transcription model Whisper has reportedly faced major issues with hallucinations, researchers told the Associated Press, something that’s particularly problematic given the model’s use in healthcare settings.
“It’s an absolute nightmare to have a medical translation tool hallucinate,” said Bob Rogers, Ph.D., the co-founder of BeeKeeperAI and CEO of Oii.ai. “The last thing you want to do is try and push out mission-critical applications and technology that’s not ready for primetime.”
But the tech in this patent (one of the very few filed by OpenAI) could be a “first step” toward making speech recognition models more robust, said Rogers. Having a one-size-fits-all approach to AI may work for some models, but for those used in critical applications, context is often key, he said. “This idea of focusing and creating tokens that control context could be a good start,” Rogers said.
Plus, a major issue with “far-ranging” models is the domino effect that can occur as it learns and grows. “You change things in one place and you get impacts in others — it’s really hard to control,” Rogers said. “Maybe focusing helps with that as well.”OpenAI wants to hear you better.
Quantum computers run hot. Google wants to cool them down.
The company is seeking to patent a “cryogenic cooling system” for “multi-unit scaling of quantum computing.” Google’s tech uses multiple stages of cooling to properly get a quantum computer to the extremely low temperature it needs to operate.
The reason quantum computers need to be kept so cool is that they rely on “superconductivity,” or conducting an electrical current without energy loss, to operate, Google said. “A challenge associated with quantum computing includes cooling quantum hardware … to a temperature at which the superconducting qubits achieve superconductivity.”
Google’s tech uses seven stages to cool these devices close to cryogenic “absolute zero” temperatures, referring to the lowest point on the Kelvin scale — nearly -460 degrees Fahrenheit.
The first stage starts above 60 kelvins, or roughly -350 degrees Fahrenheit, and the drops between temperatures get smaller at each stage, reaching around 20 millikelvins (-459 degrees Fahrenheit) by stage seven. This tiered method ensures that there isn’t a dramatic drop in temperature that could damage the quantum components.
Quantum computing faces a number of barriers to scaling. The stability of these devices is deeply impacted by their environment. The most advanced ones have only reached just over 1,000 qubits, and, as Google’s patent addresses, they need to be kept near absolute zero.
“This patent really shows that they’re trying to improve their mastery of [quantum], at least on the cooling side in terms of temperature efficiency, because you don’t want to spend too much energy cooling,” said Ashley Manraj, chief technology officer at Pvotal Technologies.
Quantum computers need to be kept in extremely isolated environments, making their usefulness and collaboration with classical computers limited. “Getting the normal computer itself next to a quantum computer ends up being a problem,” said Manraj.
“This patent is interesting because the objective seems to be to maintain efficient stages of cooling and progressively have regular computers be introduced next to the quantum workload,” he added.
While these obstacles remain significant, the advantages of this market may be underestimated, said Manraj. A quantum computer’s main strength is optimization, he said, which in theory can be applied to practically any field: economics, agriculture, logistics, government, and more. “Basically, we don’t even know what we don’t know,” he said.
Researchers are constantly making progress on this tech, said Karthee Madasamy, founder and managing partner at deep tech venture fund MFV Partners, and many have gone from focusing on physics problems to engineering ones.
“We came to the edge of Moore’s law a few years ago,” said Madasamy. “This is really the next generation of computing.”
And while very few working quantum computers exist, they’re something that Google, IBM, Amazon, and practically every hyperscaler company have their eyes on, he said. There likely won’t be one clear “winner” of quantum computing, he said, but a combined effort of “multiple architectures” that pushes the boundaries.
“Even supercomputers cannot solve many of these problems — it’ll take them thousands of years to solve a problem that can be done in a few minutes [with quantum],” said Madasamy. “I have not seen any other technologies that can leap from current trends of computing to newer things that we can achieve.”
At AWE, XPANCEO, a Dubai-based company working on smart contact lenses, showcased a few interesting prototypes of its futuristic technology. I was able to even put my eyes close to one of them and I want to tell you everything about this experience!
XPANCEO
If you’re not new to this space, you will surely rememberMojo Vision, an incredible company based in the US that was working on smart contact lenses. Unluckily, after having reached incredible milestones, like having their CEO putting a smart lens inside his eyes, the company pivoted to manufacturing microdisplays.
This is one of the most cyberpunk images I have ever seen in my life (Image by Mojo Vision)
The dream of having smart contact lenses seemed to vanish, but a new startup called XPANCEO came to the rescue. Based in Dubai, it was founded by two Eastern European guys: Roman Axelrod, a serial entrepreneur, and Valentyn S Volkov, a renowned scientist. Together they joined forces with the mission of creating smart and AR contact lenses.
I interviewed the founders of XPANCEO a few months ago, and they gave me a virtual tour of their laboratory in Dubai. From my remote tour, I had the impression that these guys seriously have the skills to be able to deliver what they have in mind. Differently than Mojo, though, they don’t have a clear roadmap in mind, yet. They are experimenting with different approaches to this hard problem, creating many different prototypes until they understand which is the best way to fulfill their vision. XPANCEO claims to be on track to unveil a fully functional smart contact lens prototype by the end of 2026.
My interview with XPANCEO, including a remote visit to its labs
At AWE, XPANCEO was showcasing a few prototypes, and for the first time, also a smart lens in which you could put your eye in.
XPANCEO smart contact lens hands-on
The most exciting moment for me at the booth of XPANCEO, apart from meeting in person Roman and Valentyn, was putting my eyes inside a contact lens made by the company. Exactly like with Mojo Vision, people couldn’t put the lenses inside their eyes, for evident safety reasons. But it was possible to grab the demo unit and put it close to the eye so that to see some visuals inside. And this was already fantastic.
Putting the demo unit of the AR contact lens close to my eyes (Thanks to Heitor Bravi for having helped me shooting this picture)
Every time I try some smart contact lens, the magic is always the same: I see it from the outside and it just seems like a normal small contact lens. Then I put it very close to the eye, I start seeing some visuals in it, and I am like WOWOWOW, THIS IS THE FUTURE. It is always very magical and very cyberpunk.
The demo unit was a hard plastic lens having a holographic display. There was an external light source in the handle of the demo device which was emitting the light in a certain way, and the lens, thanks to its holographic display, was letting me see the imagery that I was intended to see. When I put my eyes inside it, after I managed to focus on it, I was able to see some monochrome green writings, something like a simple video emulating the menu of an application that was in use. The imagery was very simple, we are not talking about the menu of the iPhone, more like the menu of a home video game in the 80s: green straight lines and texts.
A fascinating through-the-lenses video of the AR contact lens
The visuals I could see were defined enough to see the menus and read the texts. I have the sensation that Mojo Vision had better clarity, though, because I remember crisper texts.
The FOV of the XPANCEO lens was also very limited, so I could barely see a little portion of the menu interface. According to the press release I received after I visited the booth, the lens has a 30° FOV, but I can tell you that the perceived one was less. One of the reasons is that these lenses are made to show full FOV when they are worn on the eye, but of course, I could not do this test, so my perceived FOV with the lens close to the eye was probably just in the range 10-20°.
Through the lens shot of XPANCEO. The text actually looked a bit more defined than in the picture
It’s crazy that this is possible with a power consumption of just 1-3 microwatts. I mean, actually, it’s crazy that something like this is just even possible.
Other XPANCEO prototypes
The demo lens I tried is a new prototype that the company has brought to AWE and that is called “Smart Contact Lens for AR Vision”. As I’ve said, the company is working on many possible prototypes, and some of them were showcased inside some display cases in the booth.
The XPANCEO AR lens (Photo shot by Andrea Bravo)
One of the other prototypes showcased was the “Smart Contact Lens with Transparent Electronics”, which was another hard lens that featured of course some transparent electronic components, plus in the middle a one-pixel display. All of this was made possible by XPANCEO’s patent-pending ultra-thin transparent and flexible gold conductors. Valentyn Volkov explained to us that this prototype is closer to what Mojo Vision was doing, that is it has a microdisplay in the middle. He told us that this is a clear example of the experimentation stage the company is in: for the display of the lens they are trying different approaches, both the holographic display of the lens I put my eyes in and the “standard” microdisplay of this prototype.
Notice that the electronic components on the edge of the lens are semitransparent. The dot you see in the middle is actually a display (Photo shot by Andrea Bravo)
The “Smart Contact Lens for Data Reading” was instead a demonstration of “wireless data transmission capabilities integrated directly into the lens, enabling real-time interaction with data and the ability to receive information, such as biometric data, from embedded sensors” (Yes, I copied this sentence from the press release). The lens integrated some NFC technology and it was possible to put your phone close to it so that the lens made it go to DuckDuckGo and search for “XPANCEO”. I jokingly asked Valentyn if this was the future of the promotion of my blog: in the future, when I meet people at exhibitions, I randomly ask them to smash a phone into my eye to make them go to my website (I would become blind, but at least I would avoid the pain of spelling “skarredghost” all the times…). He explained to me that the NFC thing is of course a toy interaction for exhibitions to make visitors of the booth amused and the data transmission interactions imagined for the future are actually different.
This contact lens can interact with your phone via NFC (Photo shot by Andrea Bravo)
The last prototype highlighted in the booth was its “Contact Lens with Intraocular Pressure (IOP) Sensor”. This lens featured a small square which was a sensor for intraocular pressure, which is an important parameter for the early detection of glaucoma. This sensor installed directly on the eye can constantly monitor IOP with great accuracy, and if this data is collected e.g. by a smartphone, it could be used to detect some possible health issues before it is too late. What was amazing about this lens is that it was not a hard lens, but a soft hydrogel one. This is relevant because many people who wear contact lenses every day don’t wear hard plastic lenses, but this kind of soft ones. So if XPANCEO manages to install its technology into soft lenses, it could really pave the way to the future of technology. This is an edge the company has over Mojo Vision, which only operated on hard scleral lenses.
Soft hydrogel lens with IOP sensor installed on it
Final considerations
Selfie at the XPANCEO booth. From the left to the right: Valentyn S Volkov, me, and Andrea Bravo (who visited the booth with me and helped with the photos). On the back, there is Roman Axelrod photobombing
I’m a big fan of smart contact lenses and at the booth of XPANCEO, I looked like a little kid inside Santa Claus’s village. It was very interesting to see the different prototypes the company is working on, and especially it was magical to put my eyes inside one of them. There is still a lot of work to do to have commercially available smart contact lenses, but it is great to see that there are already tangible experiments in this sense. XPANCEO is proving with these showcases that they have the capability of actually delivering something interesting.
The dream of having smart contact lenses in this decade is one we can have with our eyes opened…
Voici un petit outil bien pratique pour enregistrer votre écran sous Linux sans faire cramer votre CPU ! Ça s’appelle GPU Screen Recorder et ça utilise quasi exclusivement la puissance de votre carte graphique (GPU) pour enregistrer votre écran, en réduisant au maximum l’impact sur les performances de votre système.
Plus de processeur qui gargouille à mort dès que vous lancez une capture vidéo et cela que vous ayez une carte NVIDIA, AMD ou Intel, ce sera votre fidèle GPU qui se tapera tout le boulot, un peu comme ce fameux ShadowPlay sous Windows pour ceux qui connaissent.
Le présumé pirate de FREE affirme avoir vendu la base de données clients de l'opérateur pour la coquette somme de 175 000 $. Et la marmotte, elle met le chocolat dans le papier alu ?...
Researchers have found new versions of a sophisticated Android financial-fraud Trojan that’s notable for its ability to intercept calls a victim tries to place to customer-support personnel of their banks.
FakeCall first came to public attention in 2022, when researchers from security firm Kaspersky reported that the malicious app wasn’t your average banking Trojan. Besides containing the usual capabilities for stealing account credentials, FakeCall could reroute voice calls to numbers controlled by the attackers.
A strategic evolution
The malware, available on websites masquerading as Google Play, could also simulate incoming calls from bank employees. The intention of the novel feature was to provide reassurances to victims that nothing was amiss and to more effectively trick them into divulging account credentials by having the social-engineering come from a live human.
Free a été victime d'une attaque informatique qui a provoqué une fuite de données personnelles et de coordonnées bancaires (IBAN). Il est possible de mettre en place des actions pour détecter et parer d'éventuels prélèvements frauduleux.
We hate to admit it, but whenever we see an article about either Voyager spacecraft, our thoughts immediately turn to worst-case scenarios. One of these days, we’ll be forced to write obituaries for the plucky interstellar travelers, but today is not that day, even with news of yet another issue aboard Voyager 1 that threatens its ability to communicate with Earth.
According to NASA, the current problem began on October 16 when controllers sent a command to turn on one of the spacecraft’s heaters. Voyager 1, nearly a light-day distant from Earth, failed to respond as expected 46 hours later. After some searching, controllers picked up the spacecraft’s X-band downlink signal but at a much lower power than expected. This indicated that the spacecraft had gone into fault protection mode, likely in response to the command to turn on the heater. A day later, Voyager 1 stopped communicating altogether, suggesting that further fault protection trips disabled the powerful X-band transmitter and switched to the lower-powered S-band downlink.
This was potentially mission-ending; the S-band downlink had last been used in 1981 when the probe was still well within the confines of the solar system, and the fear was that the Deep Space Network would not be able to find the weak signal. But find it they did, and on October 22 they sent a command to confirm S-band communications. At this point, controllers can still receive engineering data and command the craft, but it remains to be seen what can be done to restore full communications. They haven’t tried to turn the X-band transmitter back on yet, wisely preferring to further evaluate what caused the fault protection error that kicked this whole thing off before committing to a step like that.
Following Voyager news these days feels a little morbid, like a death watch on an aging celebrity. Here’s hoping that this story turns out to have a happy ending and that we can push the inevitable off for another few years. While we wait, if you want to know a little more about the Voyager comms system, we’ve got a deep dive that should get you going.
The task of moving engine covers “between supplier containers and a mobile sequencing dolly” isn’t especially exciting, but it demonstrates several of the new robot’s capabilities working together. After being provided with only a “list of bin locations to move parts between,” Atlas uses various sensors and machine learning models to determine the location of bins and how it needs to manipulate its body, arms, and three-fingered...
![[miko_tarik] wearing diy AR goggles in futuristic setting](https://hackaday.com/wp-content/uploads/2024/11/diy-ar-goggles-1200.jpg?w=800)
![[miko_tarik] wearing diy AR goggles](https://hackaday.com/wp-content/uploads/2024/11/diy-ar-goggles-smallphoto.jpg?w=400)
Creating Zero wasn’t simple. From designing the frame in Tinkercad to experimenting with transparent PETG to print lenses (ultimately switching to resin-cast lenses), [mi_kotalik] faced plenty of challenges. By customizing SPI displays and optimizing them to 60 FPS, he achieved an impressive level of real-time responsiveness, allowing him to explore AR interactions like never before. While the Raspberry Pi Zero’s power is limited, [mi_kotalik] is already planning a V2 with a Compute Module 4 to enable 3D rendering, GPS, and spatial tracking.
