This has to be the fastest filmmaking hack I've ever found. And all it took was a relatively cheap gimbal, a wireless mic, and an iPhone. Seriously. No editing tricks, no ultra-expensive gear. I've managed to create a kit that targets the two biggest giveaways of amateur filming: shaky cams and poor audio.
So, for one month and counting, I've been filming everything from product close-ups to run-and-gun vlogs on my iPhone 17 Pro Max. I paired with the DJI Osmo Mobile 8 gimbal and the DJI Mic Mini.
And it's completely changed how that footage looks and sounds. The results are so good, I'm never going back. If you're videographer or creator at any level, I recommend checking out this setup.
Why the Osmo Mobile 8 makes iPhone footage look intentional(Image credit: Future)The Osmo Mobile 8 is one of DJI's smartphone gimbals, and the headline upgrade over its predecessor is 360-degree pan rotation. That means you can track a subject through a full 360-degree spin without hitting a mechanical dead-end mid-take. It seems incremental on paper, but eliminating that rotational hard stop saves you from constantly having to pause and reset your framing during walk-and-talk shots.
For iOS shooters, the real highlight is direct Apple DockKit integration. It hands subject tracking over to Apple's own framework and works across more than 200 iOS apps rather than being locked to DJI's Mimo app. Point the camera at yourself, tap to lock on, and the gimbal keeps you centered while you move around a room.
The magnetic phone clamp holds the iPhone 17 Pro Max securely despite its size, and the built-in tripod and extendable rod mean I've stopped carrying a separate stick for low, sweeping shots or basic selfie-stick framing.
The multifunctional module clips to the top of the gimbal and acts as the setup's central hub, handling three jobs at once: fill light, tracking sensor, wireless mic receiver. It can be mounted facing forward or reversed, depending on whether you're filming someone else or yourself.
The Mic Mini keeps a low profile(Image credit: Future)The DJI Mic Mini is a tiny wireless microphone system: the transmitter weighs about as much as a couple of paperclips and clips or magnet-mounts to a shirt collar without pulling the fabric down.
It records 24-bit, 48kHz audio that punches well above its size class, and its two-level noise cancellation does a solid job of cutting background hum without making voices sound processed.
Setting it up is genuinely simple. The Mic Mini transmitter pairs directly with the multifunctional module on the Osmo Mobile 8 through OsmoAudio, so there's no separate receiver dongle hanging off your phone. Audio and video are captured together, in sync, from a single rig.
Battery life is rated at roughly 11.5 hours on the transmitter, and the charging case extends that to about 48 hours total, so it's rarely the component that runs out first in my experience.
Why they're better together, and still useful apartClipped onto the gimbal, the pair works as one cohesive system: the Mic Mini feeds audio straight into the recording alongside DockKit-tracked video, and small indicator lights on the module confirm the mic is connected without needing to check an app. That's the setup I use for anything scripted, since it means one rig, one battery to monitor, and zero trailing cables.
But the Mic Mini doesn't need the gimbal to be useful. Its separate mobile receiver plugs into the iPhone's USB-C port directly, so I can use it for a sit-down interview or a desk-based talking-head video with the gimbal left at home.
It also pairs over Bluetooth straight to a phone, though that route caps out at a lower sample rate than the wired or OsmoAudio connection, so I only reach for it when convenience matters more than audio fidelity.
Real-world friction and useful tipsThe Osmo Mobile 8 relies heavily on the DJI Mimo app to unlock its full feature set, and there's a learning curve before the gimbal feels natural rather than fiddly. It's also bulkier than a pocket-sized alternative like DJI's Osmo Pocket line, so it's not something you toss in a jacket pocket on the way out the door.
That said, setup friction on the iPhone 17 Pro Max is minimal, as its USB-C port means charging and wired audio just work without requiring adapters.
A couple of practical tricks worth knowing: the gimbal features a hidden "Keep Composition" behavior. If you manually nudge a tracked subject off-center—say into the left third of the frame—it remembers that offset and holds that composition on future tracking shots rather than forcing the subject back to center.
And if you ever need to unpair the gimbal from a phone without navigating Bluetooth menus, holding the shutter button, switch button, and trigger simultaneously for two seconds forces a hard reset of the connection.
A month in, this is the fastest upgrade I've made to my footage — faster than any editing trick. Fix the shake and the audio at the source, and the clip just looks like it came from a real camera.
Take your video shooting to a new level DJI Osmo Mobile 8
The DJI Osmo Mobile 8 brings a genuinely useful upgrade to smartphone stabilization, headlined by a pan axis that now spins a full 360 degrees, giving sweeping panoramic shots and unbroken tracking motion that older Osmo models couldn't manage. Setup is effortless thanks to a magnetic phone clamp paired with a built-in sensor that automatically detects the phone and balances the gimbal, and tracking has been upgraded too: the new ActiveTrack system follows human or pet subjects automatically, keeping them in frame even with zooming and gimbal movement after they've left the shot.
Read our DJI Osmo Mobile 8 review.
DJI Mic Mini (2TX + 1RX) and Charging Case
The DJI Mic Mini packs a lot of pro capability into a tiny footprint, built for creators who want great audio without carrying extra gear. The transmitter weighs just 10 grams, offering a lighter and more discreet wearing experience than the DJI Mic 2, with both magnetic and clip-on styles so it won't weigh down a collar or flip outward on thin fabric. Battery life is a standout: the transmitter and receiver run for roughly 11.5 and 10.5 hours respectively, and a fully charged charging case adds about 3.6 more charges for a total of 48 hours of use — ideal for long interviews, travel, or all-day shoots.
Read our DJI Mic Mini review.
Now that AI vendors have us hooked and using generative and agentic tools, pricing models are shifting. Subscriptions are no longer loss-making models designed to attract experimental customers, because they now need to serve a much more important purpose – make back money.
How this might look is still largely being figured out, but already we're seeing major software vendors ditch per-seat pricing in favor of token consumption or outcomes.
What this is likely to mean for corporate and enterprise customers is that AI is about to become a whole lot more expensive – but it's not all bad news because these companies also have the opportunity to gain more control over certain spends.
Step forward the AI PC, which is increasingly capable of running small models locally and handling basic and even some mid-level generative tasks without ever needing to spend a cent in the cloud. A few months ago, my social media was even filled with consumers and knowledge workers grabbing Mac minis to run Openclaw's AI agent locally.
A brief overview of how the consumption-based model could lookShashi Upadhyay, Zendesk's President for Products, Engineering and AI, explained the reason behind the shift we're seeing in pricing models – traditional metrics are becoming less relevant and new metrics are emerging. "We believe software value should align directly with customer success, not headcount," he told me in an interview.
Upadhyay also criticized per-seat models for charging customers for raw AI, whether their problem gets solved or not. But a shift to outcome-based pricing relies on the vendor and customer agreeing "on the exact result that triggers payment."
On the surface, it looks like this could be a variable that might differ on a vendor-by-vendor basis, but thinking about it more deeply, customers would ultimately be able to determine their own meaningful outcome, meaning that they only ever pay for success and never pay for failures.
Gartner Research Director Ranjit Atwal even implied that AI PCs could follow a similar trajectory: "The real question isn’t about definitions—it’s about how PCs deliver AI outcomes." Analyst projections have AI PCs (equipped with NPUs) accounting for 95% of the market by 2029.
Distinguishing the needs of local processing vs. cloud computeWith this emerging hybrid split of use cases in mind, it would mean enterprises can pay a one-time set price for an AI PC and workers will have unlimited access to on-device processing for tasks like summarization, transcription, image background removal and more.
It leaves cloud – the more unpredictable expense – only for the tasks that require extra compute, such as huge enterprise applications that handle centralized data, or running the latest and most powerful models.
I set out to determine what exactly makes an AI PC and how analysts expect this shift to impact the relevant markets, but Omdia’s research director for PC and tablet research Ishan Dutt warned me that we're still living through this transition, so quite how the end looks is yet to be determined.
For example, Dutt explained that so-called 'AI-ready' PCs just 18 months ago would've had sub-10 TOPS NPUs. But then came along Microsoft's own classification of Copilot+ PCs with around 40 TOPS. Just recently at CES 2025, Intel, AMD and Qualcomm all showed 50+ TOPS NPUs, and we're already seeing hints of 75+ TOPS.
Market intelligence firms like Omdia would typically classify an AI PC as one possessing a Neural Processing Unit (NPU), "designed for running AI workloads locally alongside the CPU/GPU," Dutt told me, but this is clearly a new category whose ceiling is still being pushed by chipmakers, and whose baseline is still being written.
Other, more traditional metrics are also relevant in the world of local processing, with Omdia implying that 16GB or memory is barely sufficient these days. Anything more than lightweight tasks is more likely to benefit from 32GB+.
"In practice, 'AI-capability' is really a function of three hardware considerations (NPU TOPS, RAM, and increasingly GPU for generative/creative workloads), not a single number," Dutt told me in an exclusive interview. "I’d expect the bar to keep moving as agentic, always-on background AI workloads become the reference use case rather than chat-style assistants."
AI PC demand vs. the effects of the device refresh cycleHaving determined the parameters that broadly define an AI PC, I wanted to understand whether enterprises are actually kitting staff out with them today. I've read many a study about the increasing unpredictability of cloud-based AI pricing, and data even shows that AI PC shipments are rising. But is this indicative of a shift away from cloud-exclusive processing, or is it just that more PCs now classify as AI PCs anyway, and generic refresh cycles are impacting these numbers with false positives?
"NPUs are now the baseline across new Intel, AMD and Qualcomm platforms," Dutt said, before reminding that me that Macs have had NPUs since the 2020 move to Apple silicon, which came around two years before ChatGPT went public.
Omdia's market analysis projects that the AI PC (equipped with an NPU) market share among all PCs will reach 77.5% by 2030, compared with 17.3% in 2024. Analysts also note that the Windows 10 end of life concentrated a large number of upgrades over the past year or two, bringing AI PC market share up considerably.
Gartner's Atwal is more optimistic that organizations will focus more on PCs that "play a strategic role in AI workflows and business outcomes," indicating there will be at least some thought into buying higher-end devices.
Dutt explained that "clearer outlining of how hardware upgrades unlock future on-device AI functionality" could further impact the cadence of refresh cycles, while Atwal sees the greatest success coming from OEMs who can help "alleviate the token crisis" by connecting cloud to edge.
AI PCs have a pricing problem of their own"Consumption-based cloud AI pricing does create a genuine total cost of ownership argument for shifting inference workloads on-device," Dutt agreed, however the PC market has its own challenges that are also leasing to pricing instability.
While it may seem like we've been living through a chip shortage for half a decade, we've actually been in two – and this second one is probably much worse.
The earlier 2020-2022 shortage was a "pandemic-driven demand/logistics mismatch," which normalized pretty quickly. This second shortage is a "deliberate capacity reallocation as memory makers are shifting DRAM/NAND wafer capacity toward HBM and high-capacity DDR5," Dutt told me. In other words, the problem we're looking to solve (AI-driving pricing instability) is being driven by itself (enterprise AI consumption). A Catch-22, if you like.
With Apple recently doing the unthinkable and raising prices across most Mac models, or cutting entry-level models altogether, Omdia sets out the challenges that lay ahead: 60% of channel partners expect to experience either delays (53%) or outright cancellations (7%) of their customers’ hardware refresh plans.
Already, 70% of channel partners are expecting major shortages across some (11%) or most (59%) of their PC hardware products.
Through 2026-2027, Gartner is preparing for reduced sales, but 2028 could mark an inflection point where memory prices drop and agentic AI matures, pushing forward a new replacement cycle.
The future is hybridReturning to my original theory that many organizational AI workloads will shift to local processing amid an evolving cloud subscription landscape, it's increasingly clear that a rise in AI PC shipments doesn't necessarily correlate with genuine customer demand.
While shifting some processing to run locally has huge merits when looking at the cloud compute market in isolation, the PC market's own challenges complicate strategies further. The answer will likely involve a hybrid approach to both sides, but how that hybrid and split will look will vary case-by-case.
But more crucially, cloud processing isn't set to go anywhere, and it will continue to complement local workloads with the type of power and performance that few organizations could ever afford to acquire. Upadhyay reminded me that cloud compute's benefits span "centralizing multi-channel customer data, real-time ticket routing, unified reporting, and seamless API integrations across hundreds of tools."
If someone tells you a song is important to them, you should listen to it carefully. Because afterwards, you'll know them better. If they share a song with you, treat it with the kind of care and respect you'd give to a treasured photo or keepsake. It's a window into their very private world. And it is all we really have, in the end.
I knew the songs that were truly important to my mom: Eric Clapton (well, Derek and the Dominos) and Layla; Cream's Strange Brew; Crosby, Stills, Nash & Young's Sanibel (which I can never ever listen to without crying, ever again); Hendrix's Are You Experienced?; Jethro Tull's Aqualung and so many more that were on heavy rotation in the car, whenever she'd drive me to ballet class as a child.
And I knew the musicians she respected: Janis Joplin, Prince ("He's a better musician than anyone gives him credit for, Becky"), The Hollies, Led Zeppelin, Queen, Bowie, Rod Stewart, The Police, Genesis, Springsteen, The Eagles, Fela Kuti (my parents actually met in Nigeria, but that's another story)… it's a long list.
What I didn't know was that my mom had since learned to love Sam Fender, Lewis Capaldi, The Weeknd, Bruno Mars and Ed Sheeran. I only know this because she'd bought their CDs from thrift stores. I even now know a couple of the volunteers from said thrift stores. They worked out who I was from my name, when I went to donate a few things (it's so hard to let anything go when someone dies) and they asked after her.
It turns out that they used to put certain CDs aside, ready for when she came in. And apparently she always bought whatever they recommended — my mom was a kind, giving and inquisitive person.
FutureFutureI also didn't know my mom had an extra cupboard stuffed solely with CDs. The entire back catalogs of The Beatles, Jackson Browne and Jethro Tull weren't a surprise, but Wu-Tang Clan's 1993 epic Enter the Wu-Tang (36 Chambers) was likely a thrift store revelation — I can even see she paid £2 for it — and I love that for her.
I should explain: my mom and I were very close. We also lost my beautiful younger sister in 2021 after an illness (not helped by that whole pandemic thing) which made us cling to each other even more. But I lived and worked in London; she lived a three-hour drive away. Distance is annoying.
You may have guessed that my mom has passed. And you may also have guessed that it was sudden, so I never got the chance to say goodbye. I never got to do it the way you always imagine you'll be able to at the end; to hug her, hold her hand, thank her for being my awesome mom and hear whatever she might want to tell me.
If you're reading this and thinking, 'Me too', I urge you to sit with your person's music for a while. And this is especially pertinent if it's music they kept in their car.
FutureFutureTangled up in blueBehold! The car my mom part-exchanged her bigger, sportier Mazda for in 2013. I coated her off for getting an "old lady car" with its "toaster engine" and begged her to reconsider, but the deal was done.
And now that this vehicle has sadly come to me, I made a curious revelation: the CD player I've always known to be in this car wasn't there when she bought it. She must have had it fitted later. How do I know? I found the original central dash column in the spare wardrobe. And she clearly loved this (then) new souped up player, because every available compartment, door and nook in this car is stuffed full of CD books — those ones where you do away with the case for space and fit the holographic discs into the little plastic sleeves.
(Image credit: Future)I love that she did this as a gift to herself and didn't tell anyone. When I first stuck the key in the ignition, Hootie & the Blowfish's Cracked Rear View was playing. We went to see them in the early noughties, at the House of Blues in Florida. I cried for a full hour at the wheel, still in the parking lot. I needed it. Actually, it felt good.
(Image credit: Future)Also in the car was Santana (we saw Carlos et al at the Hammersmith Apollo in 2016), Bon Jovi ('95, Wembley), Bruce Springsteen (Ricoh Arena, Coventry, also 2016) and several CDs I'd burned for her between 2002 and 2005. I always tried to write in a circular fashion with my black Sharpie when I created these discs, to echo the grooves of a record. There's a whole extra feature to be written about the lost art of burning a CD for someone you love, I think.
(Image credit: Future)Oye Como VaIt wasn't just acts we'd seen live taking up the limited space in her car, of course. My mom kept a lot of Bob Dylan around; Blood on the Tracks was clearly the last thing she'd been spinning before switching the disc out for a bit of Hootie — and Traveling Wilburys' Vol. 1 was just below that, in the little cubby hole above the CD player.
In the time since I lost my mom, I've driven this car a lot. I have often driven it at night, when I couldn't sleep. On the road, I belted out the songs I burned for her back when that was a thing you did for people you loved. It was my grief therapy. I shared playlists with her friends and we talked about gigs they'd been to with her. They say it helped them too.
If I couldn't make a decision about what she'd want me to do about things, I put one of her tracks on in the car and found the answer. And it was invariably, "Don't worry about it. Live your life and be free, Becky. I love you".
I hope she knows I love her too. I miss her. And I definitely approve of her taste in music.
(Image credit: Future)Music has always been a non-negotiable in my life. I see now that it was a non-negotiable in my mom's life too. She just showed it quietly and in different ways, because she was a mom to two daughters and that took it out of her.
I truly believe humans need music. We need real music with real lyrics, that actual people had to write down or they couldn't rest. We need the chord progressions that had to get out of their heads and onto a tape. And we need the meaning these tracks take on years down the line — their association with the lives and loves of people close to us.
Old, scratched, starting to struggle; those compact discs are like my mom's little Madza — becoming harder to run. To be honest, I know the feeling. Parts of the car are held together with gaffer tape (aren't we all?). At its last service, the intake of breath over the teeth and slow shake of the mechanic's head seemed a little graver than before.
(Image credit: Future)But I will never sell that car, or the collection of CDs within it. Not that I'd get much for either, anyway.
I have named the car Hidders, short for Hidalgo, after our gypsy heritage. To me it will always be priceless. And somehow, owning the music my mom casually kept in her ever-evolving life gig on wheels has brought me closer to her.
Now that I've listened to every CD that was squirrelled away in Hidders when I got it, I've started taking my sister's old collection of discs in there too, whenever I need to chat with her. Oh yes, we carefully kept her stacks of Erykah Badu, Lauren Hill, Prince, Maxwell, Sinéad O'Connor, D'Angelo, Red Hot Chili Peppers, Beck, Madonna, Black Eyed Peas and more too.
And don't tell me you never talk to people who may or may not physically be there in the car? You should try it, it's great.
Love you, mom and Jo. See you both down the road.
My day usually starts with a scan around major retailers, specialist sites, and deals aggregators to get a snapshot of the best tech offers that are available - and then I share the very best with you. Results can be mixed, depending on the time of year and whether there’s a major sale on, but there’s been one consistent theme lately: the deals on most PC components are atrocious.
If you’re a regular in this space, then you’re probably already well aware of the situation with RAM and storage prices over the past few months. It’s got to the point where I’ve all but given up on finding any good deals in this space, especially when there are no signs that it's improving.
Where did all the deals go?There’s no clearer evidence of just how dire the situation is with this Samsung SSD ‘deal’ that’s currently live on Amazon. This Samsung 870 EVO is advertised as a whopping 62% off. But that’s from the current list price of, um, $1,039.99. With that price cut, it’s now down to $399.99. Almost $400 for a 2TB SATA SSD? One that has been as low as $79.99 in the past? And this counts as a deal?
That’s just a snapshot of how bleak it is shopping for PC gaming components in 2026. Worst of all, there doesn’t seem to be any positive news on the horizon. You have SSD controller manufacturers saying that “the retail SSD market has almost disappeared”, RAM prices have increased by at least 400% in some regions, and now one chip maker says its production costs are set to increase by 10% next year - so CPUs are potentially next up for price hikes.
It puts into perspective just how lucky I was when I chose to upgrade my PC towards the end of 2025. There were rumblings of price increases and stock shortages in those few months towards the end of the year, but it hadn’t reached the disastrous depths we find ourselves in now for wannabe PC builders.
Here's the price history for the Samsung 870 EVO SSD at Amazon in the USFuture / KeepaAnd this is how the price of a 32GB kit of Corsair Vengeance DDR5 (CL36) RAM has changed over 12 months.Future / KeepaI did have to wait a few months for my RAM to arrive, after the manufacturer I was about to buy RAM from completely shut down, but the cost didn’t make up nearly 50% of my entire budget like it would in today’s prices.
And my PC is very much a mid-range device with a Ryzen 9600x, a 9060XT, 32GB of RAM, and a 2TB NVMe SSD, which is about all I wanted given I mostly use it for work. I do like to play the occasional game on PC that I can’t get on the PS5, or if there’s one I would prefer to play on that platform, or when prices make the PC version a more sensible choice. Those specs cover my needs nicely at a reasonable budget.
Calculating what the spend would be these days for the same spec is a terrifying exercise. It’s over £1,600 now, when I was comfortably under £1,000 at the end of last year.
What does the future hold?With everything pointing to similar or increasing prices for a long time ahead, anyone looking for a PC upgrade has no choice but to put up with these prices for the foreseeable future. Yes, there will be a chance for some savings during Black Friday in November, but we are still looking at a significant markup compared to the savings of old.
In fact, and I usually wouldn't recommend this, but with the way component prices are lately, I'd actually lean more towards buying a pre-built PC if you're shopping around now. Sure, these companies are still suffering from the same price increases, but you can actually find some good value rigs.
(Image credit: Future / iBuyPower / Edited by Gemini)Most recently, for example, the Presidents' Day and 4th of July sales at iBuyPower had some strong discounts on high-end gaming PC, with extra savings of up to $350 available on PCs with a Ryzen 7 7800X3D, an RX 9070 XT, and 32GB of RAM.
That would be my advice based on the poor quality of deals that are available on many PC components. Yes, there are still a few genuinely good offers out there that I've seen on the likes of CPUs and GPUs, but storage and RAM are certainly beyond saving right now — and I'd move quickly if there's anything good on the former, as those deals may be disappearing before long as well.
Cloudflare is set to make a major change to huge portions of the internet after standing between websites and automated traffic in a bid to protect publishers and hosters. Soon, the company will divide AI-related traffic into one of three categories – search, agent or training – in a bid to give customers more control over just how much automated traffic they’re prepared to let in.
From September 15 2026, new Cloudflare customers and new sites created by existing customers will block agent and training crawlers by default on pages that display ads, arguing that publishers need human visitors to see those ads to generate revenue - with customers able to also change these settings at any time through their dashboard.
But huge challenges stand between now and such a system being rolled out effectively, because many major crawlers like Googlebot are defined as multipurpose and don’t exactly fit into one of those three categories. In the worst case scenario, this could prevent any automated traffic from reaching a site, even if some categories are permitted.
Cloudflare vs. Google: A battle in which publishers could be the actual losersWorryingly, Googlebot specifically plays a key role in discovering and updating pages in Google Search, and preventing its access could have major implications on content indexing.
Pages that reject Googlebot on the basis that publishers wanted to block at least one of those three categories could gradually lose search visibility, leading to major commercial implications.
Despite the clear disadvantages of blocking automated traffic, many assert that blocking crawlers helps to protect their intellectual property, reduces web hosting infrastructure demands caused by unpredictable traffic spikes, and gives owners more control over how and where their information is used.
Varn Search Marketing Search & Innovation Director Andy Mollison describes the situation as a “high-stakes staring contest” between Cloudflare and Google, but more importantly, one in which publishers are actually the ones who pull the short straw.
I spoke with Mollison to understand how publishers and businesses should move forward to strike the right balance between protecting their content and remaining visible online.
Cloudflare and Google are currently having a high-stakes staring contest. On paper this looks like a mismatch. And in reality, it doesn’t look like it will work out for Cloudflare.
There is a simple reason Google isn’t going to blink first; the only losers from this will be Cloudflare’s customers. Mainly, publishers.
People will still be using Google, blissfully unaware a significant portion of the web is no longer included in search results. Cloudflare customers that rely on ad revenue, on the other hand will see their traffic drop off a cliff.
The goal is potentially admirable; allowing publishers to protect content from being stolen by AI, but by setting the default to ‘block all AI bots’ publishers that aren’t following every move Cloudflare makes closely will be the ones who pay the price.
The worst case scenario is every website with ads currently using Cloudflare will be completely blocked from Google search.
This obviously poses a problem for publishers in terms of the organic traffic loss. The knock on effect is then decreased ad revenue. Similarly for B2B businesses and consumer brands, this lack of search visibility will harm their bottom lines.
Consumers may not initially notice, but eventually they might question why results from publishers stop appearing.
It is the answer to why most businesses do most things, but the simple answer is money.
Cloudflare’s CEO declared last month that for the first time in the internet’s history, machines now generate more web traffic than people. AI crawlers are a big reason why. They take up a lot of crawl budget on servers, and therefore Cloudflare needs more resources to handle this increased traffic.
Cost is probably the primary reason, but a secondary aspect is autonomy. Cloudflare wants to give content publishers the ability to decide what level of indexation their content receives, depending on the intention of the AI robot.
For example, you could block your content from being put into an AI's training model, but still keep it available for when someone performs a 'search' in an AI platform.
This gives publishers better choice on where their content is used in AI platforms, on a page by page basis.
A little bit technical here, but bear with me. Cloudflare has been quite persistent with trying to recommend an llms.txt website file, which could in theory be used to dictate which parts of your website are open to AI crawlers. And more importantly, which parts are not.
Google's stance has been to state that an llms.txt file is not something they recommend in their Developer Fundamentals guidance. However, they have also created pages in their Agentic Browsing documentation about how to set up an llms.txt file.
So the confusion comes in Google stating they don't use llms.txt files, yet still creating documentation for using it.
There feels like there is a misalignment between Google's developers, which in itself then becomes difficult for website owners to know which stance to take. Either Google can crawl without use of the file, or they need it. It can’t be both.
A significant portion of searches now ends without a click. For publishers this means they are seeing their traffic dramatically slashed as a result. This then has a knock on impact on advert revenue and affiliate deals. Therefore they might want to avoid their content showing up in the AI overview, in theory, hoping users would have to come to their website to read the stories.
The likelihood however is probably that consumers simply would continue to use AI search, and their information would come from other sources instead. This means blocking an AI crawler is potentially even more harmful.
Not at the moment, other providers have kept very quiet about the situation as a whole. It feels like Cloudflare is using AI search as an opportunity to make noise and stand out from the crowd. Whether it works in their favour, we'll have to wait and see.