News
- TikTok is releasing a new tool to help users identify AI spam
- It says that AI spam “crowds out original creators”
- Industry insider suspects that ad revenue may be the real motivator
If you're getting tired of AI slop on social media, you're not alone. Even TikTok, which has spent the last couple of years rolling out AI creation tools, is now testing new ways to detect AI-generated spam and teaching users how to recognize AI content.
TikTok says it is testing improved detection systems to target accounts dedicated to posting AI-generated spam that “crowds out original creators”, and it is also creating a new guide that helps its users use AI tools responsibly. TikTok is launching a new in-app hub within the next few weeks where you can learn practical skills for spotting AI-generated content when you search for AI-related terms.
(Image credit: Future)Is it really a moral stance?We spoke to Donatas Smailys — CEO of Billo, the largest UGC creator marketing platform in the US — about the new tool. His take is that TikTok is fighting AI content because it doesn't perform, and that hurts TikTok's ad business too.
"It would be naïve to read this as TikTok taking a moral stance. The platforms can see in their own data what audiences respond to, and they're quietly rebuilding their rules around real people.”
It strikes me as slightly ironic that TikTok, which has spent the last couple of years pushing AI creation tools, is now simultaneously educating people to recognize AI content and testing systems to suppress AI spam.
If TikTok's own data suggests AI-generated spam is harming engagement or advertising performance, then it may be the saving grace for the rest of us who are tired of AI slop ruining the experience of social media. Frankly, at this point I’ll take anything that reduces the amount of AI videos that get served to me, for whatever reason.
What this means for the futureThe wider issue here is that if TikTok is beginning to treat low-quality AI spam as a platform problem, then it should raise some red flags for advertisers who heavily rely on synthetic influencers and AI creators. If low-quality AI spam was improving the platform, TikTok would have little incentive to invest in detecting and suppressing it.
TikTok actually has pretty stringent rules on AI content on its platform already. It’s quite common to see the label “Contains AI-generated content” on TikTok videos.
It seems that regulators are moving in the same direction. “New York now requires disclosure of synthetic performers in ads”, says Smailys “with penalties for those who don't comply. And my guess is that more states will follow. But the loss of trust is an even bigger risk here. Once your audience feels tricked, no disclosure label wins them back."
While the rise of AI slop seems unstoppable at the moment, I wonder whether we'll eventually look back at 2026 as the year platforms realized that flooding people's feeds with AI-generated content wasn't sustainable and that what users really wanted all along was more real people.
Quantum computing has fascinated the tech world for several decades. That’s because, while classical computing is binary, quantum bits (qubits), made from atoms, electrons and superconductors, can be in both “0” and “1” at the same time, providing computational power that is inaccessible to classical computers.
The implications of this computational power will be far-reaching, such as accelerating drug discovery, optimizing global supply chains, and revolutionizing material science for clean energy. If that’s not interesting enough, the new frontier of hybridizing quantum and classical computing generates yet another source of excitement for the field.
Humanity has embedded computing into the fundamental fabric of daily life. It underpins our progress in an enormous number of fields, as well as impacts our day-to-day lives more than almost anything else.
While quantum computers are fascinating on their own, it is becoming increasingly clear that their integration into data centers - treating them as a crucial piece of the broader computing puzzle - represents one of the most interesting frontiers in science and technology today.
Hybrid computing: already a reality and key in the futureThe idea that quantum and classical computing will work together is not a future vision. Quantum computers already rely on tight integration with classical computing, which plays a critical role in enabling their operation.
For example, extensive classical computing is used for generating the control signals that operate the quantum hardware and to calibrate quantum hardware and control systems. In this scenario, classical computers are finding thousands of different system parameters and retuning them constantly as they drift over time.
Very compute-intensive classical algorithms are also used for decoding errors during quantum error correction, which is critical for enabling stable quantum computation and the scaling of quantum systems. Another key use case is that hybrid quantum-classical algorithms rely on interleaving quantum and classical computation to solve complex problems.
In the future, this hybridization will have a growing impact on computing applications. QPUs will act as specialized accelerators within classical HPC and AI workflows, complementing CPUs and GPUs rather than replacing them. Quantum systems will sit alongside CPUs and GPUs in data centers and be deployed for exactly the type of problems they are best suited to solve.
In fact, this is already starting to happen - some of the most important recent quantum demonstrations have relied on quantum and classical systems operating in close coordination. For example, RIKEN and IBM scientists recently achieved one of the largest quantum simulations of iron-sulfur clusters through closed-loop data exchange between a co-located IBM Quantum Heron processor and RIKEN's Fugaku supercomputer.
This kind of hybrid exchange between quantum and classical systems, rather than quantum computing operating as an isolated, standalone resource, is the model we expect to become standard as the technology matures.
The final, and perhaps most intriguing frontier in quantum-classical hybridization is its relationship to Artificial Intelligence. Compute infrastructure is being used to train AI models, while hardware capabilities limit what models can do in various ways. Integrating quantum hardware into this compute fabric may allow new data to be generated for AI models as well as new engines for AI inference.
Moreover, AI tools may help us break some of the abstraction layers we have created, which can remove further limitations on how we use quantum hardware to solve problems.
Quantum computers don't run themselvesGiven the above, as quantum computing moves from laboratory demonstrations toward practical applications, success will depend not only on better quantum processors but on how effectively quantum and classical computing work together.
The latency and bandwidth of the quantum-classical links will play a critical role in how performant this integration is and will determine how deeply connected the quantum and classical processing can become. This matters because qubits hold their state only briefly - the faster the classical system can read, process and respond, the more complex a calculation it can support before that state is lost.
Another critical aspect is software. Building infrastructure, tools, compilers and applications that orchestrate hybrid quantum-classical workflows efficiently, with high performance as well as developer experience in mind, will be critical for the industry to succeed in its mission to deliver real-world value in a timely manner.
This requires investment in the orchestration layers that allow developers to calibrate and control quantum resources as easily as they would HPC today, abstracting away much of the underlying hardware complexity.
The next phase of quantum computing will depend on bringing together advances in physics, engineering and computer science to cross the barrier from lab prototypes to useful computers.
There are many challenges ahead, from scaling the hardware and performing error correction at scale to finding the right applications to push for. The clear message is that hybrid quantum-classical systems are not a temporary stage in this journey. They are the foundation on which practical quantum computing will be built.
Enterprises and IT leaders exploring quantum shouldn’t focus on the number of qubits, but on how systems integrate with classical infrastructure. When it comes to scaling useful quantum computing and implementing hybrid computing, the control systems, error correction and software orchestration will determine whether a system can deliver reliable performance and repeatable results.
We've featured the best business laptop.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
- The Jura E10 is a new premium espresso machine from the Swiss brand
- It offers hot and cold brewing, syrup-infused foam, and Light Brew mode
- It's available now for £1,425 (about $1,910 / AU$2,720)
Jura, known for its luxury espresso machines, has just launched a new coffee maker that packs a staggering array of advanced brewing features into a surprisingly compact package. The new Jura E10 offers hot and cold brewing, syrup-infusion for flavored drinks, and a new timer that means the machine can start brewing when you wake so your favorite coffee is ready and waiting the moment you wake up.
The feature set sounds very similar to that of the Jura J10 I tested last year, which is still one of the most advanced coffee makers I've ever used. It poured a perfectly-extracted espresso, and its cold-brewed espresso was mellow and flavorful.
Its milk system was excellent as well, whether using hot or cold milk, but what impressed me most was the Sweetfoam system that mixes the syrup of your choice into the milk as it's dispensed.
Flavored coffees are usually made by pumping syrup into the cup before adding coffee and milk, which means the flavor is concentrated at the bottom. With Sweetfoam, the syrup is evenly distributed throughout the whole drink, which means you get the full flavor with every sip, and only need to use a very small quantity. You can use regular or sugar-free syrup, and during my tests I found it had the added benefit of stabilizing the milk foam so it lasts and lasts.
Perfect for small kitchensLike the J10, the new E10 offers a huge menu of hot and cold drinks, but now they're organized into a color-coded menu that makes it easier to find the one you want.
These options now include Light Brew — a new feature not available in the J10 — which uses a smaller dose of coffee and a lower brewing temperature to prepare a drink with a less intense flavor that's ready to enjoy straight away.
The coffee-timing feature means the machine will start brewing your coffee automatically exactly when you want it to — a feature you'd normally only see on drip coffee machines. However, it's worth noting that this is only available if you also pick up the Jura Wi-Fi Connect V2 module, which adds wireless connectivity to your coffee machine and costs an additional £49.95 (about $70 / AU$100).
All of these brewing functions are packed into a surprisingly slim package that will slot more easily into small kitchens:
Dimension
Jura J10
Jura E10
Width
12.5 inches / 31.8 cm
10.6 inches / 27 cm
Height
13.7 inches / 34.8 cm
13.7 inches / 34.8 cm
Depth
17.6 inches / 44.6 cm
17.6 inches / 44.6 cm
The Jura E10 is available now direct from Jura and a handful of third-party stores for £1,425 (about $1,910 / AU$2,720). It's currently only sold in Europe, but the rest of the E range is available in the US, so I'm expecting it to launch there in the coming months.
- Intel's data center business has seen tremendous growth thanks to ongoing AI demands
- Layoffs would likely target operational efficiency over division reduction
- Share prices and quarterly earnings are back to being in a good place
Intel has reportedly warned workers in its Data Center Group that another round of layoffs could be coming soon, adding to the major layoffs that already came in 2024 and 2025.
The company has not disclosed the number of jobs being eliminated, when they would leave, or any other finer details, but Oregon Live reports that the restructuring would reduce organizational complexity and allow the division to operate more quickly.
However, the changes still come as a surprise, with the Data Center Group currently one of Intel's stronger-performing businesses.
Intel set to lay off data center business workersMore broadly, warnings of sector layoffs come as more of a shock with data center hardware in hot demand and supply chains under immense pressure.
"As part of our broader strategy to become a more focused and efficient company, (the data center group) is aligning its organization to ensure it has the right roles and skills in place to position the business for long-term success," the company told Oregon Live.
New CEO Lip-Bu Tan had already warned of major workforce restructuring in a drive for organizational efficiency.
This comes against a healthy backdrop for Intel, with first-quarter revenue for the data center division up 22% year-over-year. In fact, Intel's company share price has recovered dramatically in recent months following support from the US government. From lows of under $20 throughout parts of 2025, company shares have since grown around 5x to $105.45 at the time of writing.
That growth is set to continue, with Intel set to announce its second-quarter results on July 23 2026, targeting revenue of $13.8 billion to $14.8 billion, up from $13.6 billion in its first quarter.
- RAM pricing in Germany has suddenly gotten a lot steeper
- The price of DDR5 memory is now 448% more costly than a year ago
- On top of that, TSMC is reportedly set to increase its chip production prices by up to 10% next year, which will mean more expensive CPUs
The RAM crisis appears to be getting considerably worse right now, based on memory pricing in one country anyway — and we've also had some bad news from a big chip maker.
First off, RAM pricing in Germany is what's been under the microscope courtesy of 3D Center (via Wccftech), which regularly takes the temperature of the market in the country, finding that July has seen a notable price hike.
This is for DDR5 RAM at retailers in Germany, and prices are now at an incredible 448% more costly than a year ago (before the memory crisis kicked off). Pricing had been at 440% at the start of 2026, but then dropped to 408% in March, and stayed at that level in April, before climbing back to 419% in June.
However, this latest move to 448% represents a bigger leap, and an all-time high for the German market.
As for the bleak news from the chip maker, that's TSMC which is going to raise the cost of its chip production by up to 10% next year, according to a report by Nikkei Asia (as flagged by PC Gamer).
The reason for TSMC having to hike prices is said to be the rising cost of materials and the equipment for manufacturing chips, as well as the need to fund expansion plans for more fabrication facilities (or fabs, as they're known).
Bigger premiums will be lumped onto those ordering high-end chips (high-performance computing or HPC silicon) from TSMC off the cuff (on top of existing orders), and in these cases there could be up to a 15% hike.
Analysis: hiking misery(Image credit: Shutterstock)The good news here, such as it is, is that TSMC is reportedly waiting until 2027 before implementing these price hikes. They could have come in sooner given the climate we're currently facing, although we should remember, this isn't official, and it's just based on what Nikkei Asia has heard from its sources. But it's all too believable, and indeed arguably inevitable that further chip price hikes are coming.
What this means is that the likes of processors – or anything with a CPU in it – are likely to get pricier next year, because manufacturers who are getting their chips from TSMC will, naturally enough, pass any cost increase (or most of it) on to the consumer. And TSMC supplies a lot of the tech industry's needs, Apple and Nvidia included (as well as AMD and Qualcomm).
Regarding the RAM situation in Germany, this is just one country, of course, but it's generally reflective of what a lot of chatter is saying – that the second half of 2026 is going to bring further price rises, and that next year is likely to be even worse. That's the broad theory floated by analysts and chip makers themselves in many cases, and the German market is simply backing this notion up.
Of course, there's not much that can be done about this, except perhaps to consider buying now rather than next year. Or at this point, you may want to look at Black Friday as a good upcoming opportunity to pounce on a hardware purchase like a PC or laptop, and that's the other thing to bear in mind here: how you proceed depends on what hardware you're looking at purchasing. I've discussed this recently in some depth, after Nvidia's CEO commented on the RAM crisis and how it's not going to be over until perhaps 2030.