Agregátor RSS

Kdy vlastně lze sjednat paušální platbu služeb poskytovaných v bytě nájemci a kdy to není možné? Co když byla nájemní smlouva sjednána vadně a je neplatná, má nájemce nárok na vyúčtování služeb?

Při vývoji algoritmů s fyzikálními i jinými výpočty se používají hodnoty s různými jednotkami, které odpovídají různým veličinám. Programovací jazyky s jednotkami přímo nepracují, takže je nutné využít podpůrné knihovny.

Intel na letošek chystá procesory Arrow Lake s grafikou na úrovni Meteor Lake a dále Lunar Lake, jejichž GPU bude integrovanou premiérou nové architektury Battlemage…

AI models are power hogs.

As the algorithms grow and become more complex, they’re increasingly taxing current computer chips. Multiple companies have designed chips tailored to AI to reduce power draw. But they’re all based on one fundamental rule—they use electricity.

This month, a team from Tsinghua University in China switched up the recipe. They built a neural network chip that uses light rather than electricity to run AI tasks at a fraction of the energy cost of NVIDIA’s H100, a state-of-the-art chip used to train and run AI models.

Called Taichi, the chip combines two types of light-based processing into its internal structure. Compared to previous optical chips, Taichi is far more accurate for relatively simple tasks such as recognizing hand-written numbers or other images. Unlike its predecessors, the chip can generate content too. It can make basic images in a style based on the Dutch artist Vincent van Gogh, for example, or classical musical numbers inspired by Johann Sebastian Bach.

Part of Taichi’s efficiency is due to its structure. The chip is made of multiple components called chiplets. Similar to the brain’s organization, each chiplet performs its own calculations in parallel, the results of which are then integrated with the others to reach a solution.

Faced with a challenging problem of separating images over 1,000 categories, Taichi was successful nearly 92 percent of the time, matching current chip performance, but slashing energy consumption over a thousand-fold.

For AI, “the trend of dealing with more advanced tasks [is] irreversible,” wrote the authors. “Taichi paves the way for large-scale photonic [light-based] computing,” leading to more flexible AI with lower energy costs.

Chip on the Shoulder

Today’s computer chips don’t mesh well with AI.

Part of the problem is structural. Processing and memory on traditional chips are physically separated. Shuttling data between them takes up enormous amounts of energy and time.

While efficient for solving relatively simple problems, the setup is incredibly power hungry when it comes to complex AI, like the large language models powering ChatGPT.

The main problem is how computer chips are built. Each calculation relies on transistors, which switch on or off to represent the 0s and 1s used in calculations. Engineers have dramatically shrunk transistors over the decades so they can cram ever more onto chips. But current chip technology is cruising towards a breaking point where we can’t go smaller.

Scientists have long sought to revamp current chips. One strategy inspired by the brain relies on “synapses”—the biological “dock” connecting neurons—that compute and store information at the same location. These brain-inspired, or neuromorphic, chips slash energy consumption and speed up calculations. But like current chips, they rely on electricity.

Another idea is to use a different computing mechanism altogether: light. “Photonic computing” is “attracting ever-growing attention,” wrote the authors. Rather than using electricity, it may be possible to hijack light particles to power AI at the speed of light.

Let There Be Light

Compared to electricity-based chips, light uses far less power and can simultaneously tackle multiple calculations. Tapping into these properties, scientists have built optical neural networks that use photons—particles of light—for AI chips, instead of electricity.

These chips can work two ways. In one, chips scatter light signals into engineered channels that eventually combine the rays to solve a problem. Called diffraction, these optical neural networks pack artificial neurons closely together and minimize energy costs. But they can’t be easily changed, meaning they can only work on a single, simple problem.

A different setup depends on another property of light called interference. Like ocean waves, light waves combine and cancel each other out. When inside micro-tunnels on a chip, they can collide to boost or inhibit each other—these interference patterns can be used for calculations. Chips based on interference can be easily reconfigured using a device called an interferometer. Problem is, they’re physically bulky and consume tons of energy.

Then there’s the problem of accuracy. Even in the sculpted channels often used for interference experiments, light bounces and scatters, making calculations unreliable. For a single optical neural network, the errors are tolerable. But with larger optical networks and more sophisticated problems, noise rises exponentially and becomes untenable.

This is why light-based neural networks can’t be easily scaled up. So far, they’ve only been able to solve basic tasks, such as recognizing numbers or vowels.

“Magnifying the scale of existing architectures would not proportionally improve the performances,” wrote the team.

Double Trouble

The new AI, Taichi, combined the two traits to push optical neural networks towards real-world use.

Rather than configuring a single neural network, the team used a chiplet method, which delegated different parts of a task to multiple functional blocks. Each block had its own strengths: One was set up to analyze diffraction, which could compress large amounts of data in a short period of time. Another block was embedded with interferometers to provide interference, allowing the chip to be easily reconfigured between tasks.

Compared to deep learning, Taichi took a “shallow” approach whereby the task is spread across multiple chiplets.

With standard deep learning structures, errors tend to accumulate over layers and time. This setup nips problems that come from sequential processing in the bud. When faced with a problem, Taichi distributes the workload across multiple independent clusters, making it easier to tackle larger problems with minimal errors.

The strategy paid off.

Taichi has the computational capacity of 4,256 total artificial neurons, with nearly 14 million parameters mimicking the brain connections that encode learning and memory. When sorting images into 1,000 categories, the photonic chip was nearly 92 percent accurate, comparable to “currently popular electronic neural networks,” wrote the team.

The chip also excelled in other standard AI image-recognition tests, such as identifying hand-written characters from different alphabets.

As a final test, the team challenged the photonic AI to grasp and recreate content in the style of different artists and musicians. When trained with Bach’s repertoire, the AI eventually learned the pitch and overall style of the musician. Similarly, images from van Gogh or Edvard Munch—the artist behind the famous painting, The Scream—fed into the AI allowed it to generate images in a similar style, although many looked like a toddler’s recreation.

Optical neural networks still have much further to go. But if used broadly, they could be a more energy-efficient alternative to current AI systems. Taichi is over 100 times more energy efficient than previous iterations. But the chip still requires lasers for power and data transfer units, which are hard to condense.

Next, the team is hoping to integrate readily available mini lasers and other components into a single, cohesive photonic chip. Meanwhile, they hope Taichi will “accelerate the development of more powerful optical solutions” that could eventually lead to “a new era” of powerful and energy-efficient AI.

Image Credit: spainter_vfx / Shutterstock.com

Hra MineClone2 postavena nad voxelovým herním enginem Minetest byla přejmenována na VoxeLibre.

Jak si doma nejčastěji obstaráváte filmy? Dáváte přednost streamování za předplatné, protože je to pohodlné, nebo si filmy raději kupujete, protože je chcete vlastnit a láká vás vyšší technická kvalita? Pokud kupujete digitální licence, máte nějaký preferovaný obchod jako iTunes, Google Play, ...

IDC’s latest preliminary data on the smartphone market suggests Apple’s traditionally weakest smartphone quarter might be a little weaker than usual this year as political tension gnaws away at the company.

If IDC is correct, this unravelling has lopped a few more hairs from Apple’s Big Tech scalp, with Q1 iPhone sales down as much as 9.6%. That means Samsung is once again the temporary King of the Hill, even as China’s Xiaomi also makes gains. Comparative market share only tells part of the story, of course: Apple still allegedly sold 50 million iPhones in the first quarter of 2024, according to IDC. 

Morgan Stanley has a more optimistic view. In a client note received by Computerworld, analyst Erik Woodring wrote: “Contrary to market expectations, our Greater China Tech Hardware colleagues just raised their June quarter iPhone builds.”

Specifically, Woodring tells us analysts bumped up their “iPhone build expectations by 5%, or 2 [million] units, to 39 [million] units (-5% Y/Y) citing checks with Hon Hai and reflecting strength of legacy iPhone models in emerging markets, and relative stability elsewhere.”

Apple’s weakest quarter is weak, says IDC

When it comes to the overall market, IDC has a slightly rosy outlook. “The smartphone market is emerging from the turbulence of the last two years both stronger and changed,” said Nabila Popal, research director with IDC’s Worldwide Tracker team. 

The big trend is that global political realignment is evidencing itself in a new wave of smartphone competitors. “There is a shift in power among the Top 5 companies, which will likely continue as market players adjust their strategies in a post-recovery world,” said Popal.

It is inevitable US business will be affected by international political polarization. Indeed, as things continue to unravel, it is tempting to believe the architects of division on all sides will not rest until ordinary humans are once again reduced to speaking to each other using tin cans and string.

Hopefully we can avoid that outcome.

What can Apple do?

It’s never good to see an almost 10% decline in sales of a company’s most important product, but there are other reasons for provide optimism. Not only is Apple now actively engaged in developing new business plans for a more regulated industry, it’s also practicing its next pivot to pirouette around the twin themes of AR and AI. 

Claims Apple AI will run directly on the device should translate into an accelerant for iPhone sales, particularly among privacy/security conscious consumers and enterprise professionals. But there are other people who will welcome incredibly productive smartphones capable of handling complex tasks.

The value of new markets

Accurate recognition of the true value of growing markets remains a challenge for analysts looking to enumerate potential sales data in terms of specific company achievements. There is a possibility that Apple’s continued moves to build bigger business in India and elsewhere might not yet have been accurately baked into expectations. 

However, even if the IDC data is accurate, it’s worth reflecting that Apple’s move to make iPhones in India has been met by strong gains in local share — and there may be longer legs to find. Apple’s anticipated plan for more powerful iPhones with on-device edge AI will appeal to customers in growing markets, some of whom may have almost entirely skipped personal ownership of computers. These smart devices might yet turn out to be all the computer an even greater number of consumers need. They should be capable of replacing PCs for even more tasks.

Bicycle or hype cycle?

Indeed, while there’s plenty of excitement around AI/Generative AI (genAI) across mature markets (evidenced if by nothing else by the vast number of “Get Rich Quick” scams festooned across Twitter/X), it’s plausible to think that the true liberation of human potential will come from the democratization of access to computing these things represent. This, of course, is central to Apple’s core DNA, which has always described computers as “bicycles of the mind.”

The company won’t be alone, of course. Every tech firm is running to climb aboard the AI hype machine, in part to build big market slices in advance of inevitable regulation. But for Apple, if you also factor in second user and refurbished devices and think about actual devices in use, that means hundreds of millions will gain access to these new tools in a few months for no extra cost.

In fact, as consumers choose to use their handsets longer, the only thing that really matters when it comes to smartphone sales this year is the extent to which Apple’s forthcoming AI iOS upgrade is backwards compatible. Because today’s happy customers will become repeat customers in tomorrow’s upgrade cycle. That’s how this river flows.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grilland Apple Discussions groups on MeWe.

Apple, Generative AI, iOS, iPhone, Smartphones, Vendors and Providers

Společnosti Avast Software s.r.o. byla pravomocně uložena pokuta ve výši 351 milionů Kč. Tu uložil Úřad pro ochranu osobních údajů za neoprávněné zpracování osobních údajů uživatelů jejího antivirového programu Avast a jeho rozšíření internetových prohlížečů (Browser Extensions), k čemuž docházelo prokazatelně po část roku 2019. Společnost Avast v rámci poskytování služeb antivirového software zpracovává osobní údaje uživatelů tohoto software. V prokázaném období části roku 2019 předávala část těchto údajů přibližně 100 milionů svých uživatelů společnosti Jumpshot, INC., a to zejména pseudonymizovanou historii prohlížení internetu, navázanou na jedinečný identifikátor. Společnost Jumpshot, INC. se mimo jiné prezentovala jako společnost zpřístupňující data „marketérům“, jimž poskytovala „vhled do on-line chování spotřebitelů“ a nabízela „sledování cesty uživatelů na atomární úrovni.“ Uživatelé tak byli společností Avast mylně informováni o předávání anonymních údajů za účelem analýzy trendů. Ačkoli společnost Avast uváděla, že používala robustní anonymizační techniky, bylo v řízení prokázáno, že předávané údaje z jednotlivých instalací antivirového softwaru nebyly anonymizované, jelikož i na základě předávaných dat mohlo dojít k opětovné identifikaci minimálně části subjektů údajů. Navíc účelem zpracování těchto údajů nebylo (pouze) vytváření statistických analýz, jak společnost Avast uváděla. „Úřad v rozhodnutí zdůraznil, že společnost Avast je jedním z předních odborníků na kybernetickou bezpečnost, který nabízí veřejnosti nástroje k ochraně dat a soukromí. Její zákazníci nemohli očekávat, že právě tato společnost bude předávat jejich osobní údaje, respektive údaje, na jejichž základě by mohla být zjištěna nejen jejich totožnost, ale například i zájmy, preference, bydliště, majetkové poměry, profese a další údaje týkající se jejich soukromí,“ uvedl k rozhodnutí předseda Úřadu pro ochranu osobních údajů Jiří Kaucký. Vzhledem k tomu, že se jednalo o případ přeshraničního zpracování osobních údajů klientů v rámci celé Evropské unie, byla věc řešena zároveň s ostatními dotčenými dozorovými úřady v Evropské unii, a to v rámci mechanismu spolupráce One Stop Shop.

A resurgence of cyberattacks targeting Linux systems in Asian campaigns through the utilization of the Pupy Remote Access Trojan (RAT) has been observed. The malware's multifunctional nature is a notable characteristic, striking a chord with Linux admins, infosec professionals, internet security enthusiasts, and sysadmins who are likely familiar with the potential threat of versatile malware.

A security flaw impacting the Lighttpd web server used in baseboard management controllers (BMCs) has remained unpatched by device vendors like Intel and Lenovo, new findings from Binarly reveal. While the original shortcoming was discovered and patched by the Lighttpd maintainers way back in August 2018 with version 1.4.51, the lack of a CVE identifier or an advisory meant that

A security flaw impacting the Lighttpd web server used in baseboard management controllers (BMCs) has remained unpatched by device vendors like Intel and Lenovo, new findings from Binarly reveal. While the original shortcoming was discovered and patched by the Lighttpd maintainers way back in August 2018 with version 1.4.51, the lack of a CVE identifier or an advisory meant that Newsroomhttp://www.blogger.com/profile/[email protected]

Ujíždějící vozidlo pronásledované policejními vozy není jen obrázek z akčních filmů, ale realita českých silnic. Snaha dohnat a zadržet automobil s řidičem, který odmítá zastavit, skýtá nebezpečí pro ujíždějícího, policisty i náhodné kolemjdoucí. Policie má k dispozici celou řadu prostředků: ...

Klasické tlačítkové mobily se staly okrajovou záležitostí • Někomu ale k používání stále bohatě postačují • Z nabídky prodávaných modelů vybíráme 14 nejlepších

Když dojde k havárii letadla, bývají jedním z nejdůležitějších vodítek při vyšetřování nehody letové zapisovače – takzvané černé skříňky. Ty zaznamenávají komunikaci pilotů v kokpitu i údaje o různých parametrech stroje – například nastavení tahu motorů, vysunutí klapek, letové výšce a další. V ...

Hraní šachů poháněné počítačem pokládáme za vyřešené od roku 1997, kdy Deep Blue od IBM porazil šachového velmistra Garry Kasparova. Nyní ale Google vyvinul novou AI, která umí hrát šachy. A vzhledem k původu v Googlu trochu paradoxně bez hledání. Uděláme to jednoduše Tradiční přístup k ...

Streamer says access came via credential stuffing

Streaming giant Roku is making 2FA mandatory after attackers accessed around 591,000 customer accounts earlier this year.…

Attackers could nab an org's most sensitive keys if left unaddressed

Updated  Customers of Delinea's Secret Server are being urged to upgrade their installations "immediately" after a researcher claimed a critical vulnerability could allow attackers to gain admin-level access.…