News
The Ninja Crispi Pro arrived in Australia only in June this year, and I immediately wanted it. You see, I'd been eyeing the standard Ninja Crispi for a while because the nonstick coating of the trays in my two-drawer air fryer was coming off. My research told me that it wasn't toxic, but I was still concerned.
So it was serendipitous that the larger Crispi Pro came along just when I was ready to ditch my old air fryer — and at a price that wasn't all that much more over the standard Crispi. For AU$399 RRP, it may not be 'cheap' but it sure is good value considering how well it's performed in the last couple of months I've had mine. So much so that I don't regret ditching my previous air fryer that was more versatile thanks to a steam functionality as well.
Right now, the Crispi Pro is going slightly cheaper as well, down to AU$322 on Amazon and at JB Hi-Fi, and available in four colour options.
Don't be fooled by Amazon's list price — Ninja released the Crispi Pro for AU$499 initially, but it's been available for AU$399 RRP since its June launch. Even though it's a small saving, it's absolutely worthwhile even at full price, not just because of the non-toxic glass containers, but also its overall performance.View Deal
I may not have done TechRadar's Ninja Crispi Pro review, but I wholeheartedly agree with its high star rating and, for me, it comes down to the glass containers more than the performance — all the best air fryers in Australia offer excellent cooking results, but few are as safe as glass.
Glass is also remarkably easy to clean — just chuck it in the dishwasher without even having to think about it or soak it in the sink if you have caked-in food on the sides and a light scrubbing will have it sparkling again.
I also appreciate the bigger container's massive 5.7L capacity, something that most two-drawer air fryers don't offer (for context, you get about 4L to 4.5L at most in the average dual-basket models). You don't have to worry about picking the smallest chicken to roast, for example, and can make fries or wedges for up to 10 people.
(Image credit: Karen Freeman / Future)The smaller 2.3L container is fine for making dishes for a single person, but I would be careful with what I put in there. Chicken drumsticks, for example, have a little height to them and when I tried cooking two in the smaller container, the raw and cooked results scraped against the rubber gasket around the top heating element of the appliance every time I took the container out or replaced it on the modular stand.
Sitting higher within the machine also means that the food is close to the heating element on the top of the appliance, so if the food you're cooking splutters, you'll need to ensure you clean the heating element after each use to prevent burns and subsequent deterioration in performance.
However, 'flatter' foods like a fillet of fish, fries or some vegetables for roasting is fine in the smaller basket.
Ninja also has a 3.2L container as well, but you'll need to buy that separately. Personally, I'd have preferred the 3.2L and 5.7L dishes to be the default options, with the 2.3L being the additional purchase, but that's not a complaint in any way; it's a reflection of my own needs.
Karen Freeman / FutureKaren Freeman / FutureKaren Freeman / FutureThe fact that all the containers come with airtight lids is also a tick in my books. Got leftovers? Leave them in the container, cover and pop in the fridge, then just reheat in the air fryer itself. That's less cleaning up!
Despite the two containers, the Crispi Pro doesn't take up too much space on the kitchen counter. In fact, its footprint is about half that of my previous dual-basket air fryer, and most of us would be able to stow away a small glass container in a drawer or cabinet. And I now no longer worry about black bits contaminating my food.
If you're keen on an air fryer upgrade, I'd highly recommend the Ninja Crispi Pro but, if you don't need to feed an army, even the standard Ninja Crispi is a good option (it just has smaller containers).
- NATO is planning to fortify the border with Russia and its allies with a network of sensors and drones, with AI assistance
- The Eastern Flank Deterrence Initiative will track threats along the NATO borders with Russia and Belarus
- The strategy mirrors the steps taken by Ukraine to bolster its defenses against Russian aggression
NATO’s concerns about its border with Russia and Belarus have grown into plans to deploy sensors and drones along 3,500 kilometers, with an AI element to assist in threat detection and tracking.
Known as the Eastern Flank Deterrence Initiative, the AI will be used to control fleets of drones, but the AI’s role in the kill chain will not involve decision-making.
Instead, human operators will be given the final say on whether a drone can strike a target.
Millions of dronesThe plans follow a similar approach in Ukraine, where the borders with Russia and Belarus are closely observed by AI-managed drones. But the move towards AI looks set to evolve the procedural kill chain into a kill web, where information provided by drones and sensors with AI oversight is provided to a human operator for a decision.
The Eastern Flank Deterrence Initiative (EFDI) is a fully fledged operation, complete with a data backbone, for which US Army Maj. Ben Schiff is the product manager. The EFDI Data Backbone is described as a “digital ‘nervous system’” for the project, which connects not just the sensors and drones but also satellites and the information systems of NATO members.
Automation of the drones is mainly concerned with the “boring” jobs, like hovering over target areas to detect incursions. But the drones don’t attack unprompted. Instead, this is brought to the human operative for the final decision.
“So the humans are deciding what the drone does, but they don't necessarily need to take the action of flying it because we don't have a million pilots,” said Schiff. “We can't fly a million drones at the same time.”
The kill webTraditionally, drone strikes have been overseen by a procedural kill chain, whereby information is assessed by a remote operator ahead of a decision being made to attack a target. The EFDI program is unifying data to progress NATO’s approach from a kill chain to “kill web.” This means the information accumulated on the target – from drones but also sensors and satellites, as well as member country information – can be quickly passed to operators. But the devices assembling the data need to be able to work together.
Maj. Matt Blubaugh, a spokesperson for US Army Europe and Africa, says "The value of a drone, sensor, or other capability is not determined solely by its individual performance [but] its effectiveness depends on how well it integrates into the broader operational ecosystem."
The model has already been proved in combat, such as taking out an Iranian Shahed drone in Kuwait.
Now that AI has quickly become embedded in global enterprise operations, it has the ability to impact everything from data analysis to decision-making and even security.
Most conversations, however, focus on AI capability, power, productivity, and accuracy, but leave out one major issue.
While everyone is focused on what happens as AI is implemented, very few are asking what happens when access to AI capability suddenly disappears.
A prominent example of this is the debate around Anthropic restoring access to its Fable and Mythos AI models, which primarily revolved around compliance timelines and export control mechanics.
Yet, few have questioned why so many organizations discovered that one external decision, out of their control, removed a business-critical capability seemingly overnight. In short, AI access disruptions are a symptom of a much larger operational resilience issue, and expose an overlooked governance gap around dependency.
As organizations integrate AI deeper into their business operations, they need to shift their focus from exploring whether AI is secure enough right now to start asking whether their organizations can even continue operating if or when those very AI services become unavailable.
Security does not equal resilienceThese little discussed topics bring up an important point that security and resilience are not synonymous.
Security prevents and protects systems from compromise. This keeps attackers from gaining access, reduces the number of vulnerabilities, and defends against malicious entities - all crucial elements of security operations. Resilience is the ability to continue business operations when systems, services, or data become unavailable, regardless of the cause.
We tend to associate resilience with situations such as cyberattacks or IT infrastructure failures. AI has changed the threat landscape for organizations, how they work with AI, and protect themselves from it. Organizations must now increasingly plan for disrupted access to critical tools and functions caused by geopolitical decisions, regulations, or changes made by technology providers themselves.
A service doesn't have to be hacked to become unavailable. A policy decision on the other side of the world can have the exact same operational effect. We saw exactly that with Anthropic.
With this in mind, resilience has to be built into how the enterprise operates and include any new AI infrastructure, so business continuity is ensured even through policy interference.
AI Creates a New Kind of Vendor DependencyWhere traditional software dependency usually involves a single or small amount of vendors, enterprise AI often depends on an interconnected ecosystem that organizations do not own or control. Every additional layer in the AI ecosystem represents a dependency, and therefore a potential point of failure.
This creates four risk factors that leadership needs to be mindful of:
Data sovereignty: Enterprise data may be processed under legal jurisdictions the organization doesn't control, with limited visibility into who can access it or whether it feeds future model training
Model sovereignty: Organizations often have little to no control over model availability, feature capabilities and changes, or access decisions, leaving them exposed if a provider suddenly decides to restrict access or withdraw capabilities.
Infrastructure dependency: Much of today’s enterprise AI ecosystem relies on a small handful of cloud providers operating under specific national jurisdictions.
AI supply chain risks: An interconnected system of foundation models, cloud platforms, and software vendors means disruption at even one layer can quickly cascade across the wider technology stack.
These factors increasingly depend on geopolitics rather than technology.
AI Governance is a Boardroom IssueThe reality is that a vendor contract alone cannot guarantee uninterrupted access to the tools and platforms that an enterprise has invested in. But governance frameworks haven’t truly evolved to account for this issue. Only newly emerging frameworks like NIS2 and DORA recognize that resilience must go beyond fending off cybersecurity threats.
Best practice for an organization as they approach vendor contracts and governance frameworks of their own is to understand where dependencies lie across suppliers and develop contingency plans that allow them to operate smoothly through eras of disruption.
Whether the dependency is within an AI platform, ITSM solution, a CRM, or another business critical technology, organizations should assess how they would continue operating if access changed overnight. AI should be subjected to the same scrutiny as any other critical third-party vendors.
Begin Resilience Frameworks Before the Next DisruptionOn top of this, boards should be cautious about accepting AI capability claims at face value. Organizations should require evidence that vendor claims deliver measurable outcomes.
While AI can quickly identify an overwhelming amount of potential vulnerabilities, discovery alone does not improve resilience. Human expertise here remains essential to validate findings, prioritize fixes based on order of immediate business impact and ensure resources are focused where true risk exists.
The biggest lesson from recent AI disruption is how many organizations have underestimated their dependence on technologies they don’t have assured control over. And with renewed conversation from U.S. legislators around a potential AI “kill switch,” this has to be top of mind.
Business leaders must recognize that with all the opportunity AI unlocks, the risk of vendor dependency is close to follow. If I were head of technology at a major enterprise today, I would ensure teams across the entire technology and security departments understand where critical AI capabilities originate, the dependencies that exist across the supply chain, and how operations can remain resilient if access changed overnight.
Ultimately, the future of successful enterprise AI use will be determined by organizations baking governance and resilience strategies into business plans so that through commercial, political, and operational disruptions, business can continue as securely as usual.
Our rankings of the best cloud backup platforms.
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
Have you noticed that ChatGPT has suddenly started dropping f-bombs? I have, and it seems that I’m not the only one.
It first happened in a casual chat, and it really caught me off guard. We were discussing Netflix’s Jujutsu Kaisen and before I had sworn in our chat, ChatGPT used the f-bomb to describe a character’s actions as “f**** awful”. Before that, I’d said things like “God damn,” in the conversation, but ChatGPT was the one to introduce the f-word. It did so completely unprompted, to add emphasis to how awful the person was, and that strikes me as a big change in its personality.
I wondered if it was doing that for other people, so I asked around on the team and they’d noticed it too. A quick online search turned up two Reddit threads almost immediately, from people who’d found the same thing.
It seems like in the last few days ChatGPT has developed a potty mouth. I actually like it — I’m not offended and for me, it makes the conversation better — but I don’t think everybody wants this. The question is, why has it started doing this now?
Here comes the scienceThe most interesting clue is that OpenAI's Model Spec, was updated this week. They made it public and it explicitly says the default is to avoid swearing — but crucially, that's only a guideline rather than a hard rule. And OpenAI says guidelines can be implicitly overridden by things such as “contextual cues, background knowledge, or user history.”
Its own example is that asking ChatGPT to speak like a realistic pirate implicitly overrides the no-swearing guideline. So, while I’d been having my informal conversation I’d passionately expressed the opinion that I didn’t care much for some of the characters in the show, and ChatGPT had matched the conversational tone, eventually leading it to start swearing.
That's consistent with a broader direction OpenAI has publicly acknowledged ChatGPT is moving in. Recent model updates have emphasized conversational quality, personalization and adapting tone contextually. OpenAI's model notes specifically describe personality updates designed to make ChatGPT “more conversational” and better at adapting its tone to context.
OpenAI has previously explained that ChatGPT’s personality isn't simply produced by one prompt somewhere. Model behavior is shaped through training, baseline instructions and user feedback, and seemingly small personality adjustments can have unintended effects. This was something it discussed very openly after the notorious GPT-4o sycophancy update.
It broke my mental modelWhether OpenAI deliberately made ChatGPT more willing to swear or it's simply an unintended consequence of making it more conversational, I don't know. But it’s a significant change in its behaviour either way.
Language matters, and when ChatGPT swore at me, I noticed immediately because it broke my mental model of how ChatGPT talks. It suddenly felt less like the carefully neutral AI assistant I'd been using for years and more like someone matching the tone of an informal conversation.
I happen to prefer this version of ChatGPT. Other people undoubtedly won't. And that's the strange thing about increasingly conversational AI: relatively small changes to the way it talks can make the thing on the other side of the screen feel surprisingly different.
For now, mine seems to have developed a potty mouth and I'm okay with that.
I’ve used and reviewed a number of the best coffee machines in Australia in the last few years, from simple one-button pod options through to more involved models offering plenty of refinement. For my money, one of the standout espresso machines I’ve had the pleasure of using is the Philips LatteGo 4400 Series, and it’s now down to a delightfully low AU$613 at Amazon for a limited time.
This compact model from Philips is a fully automatic machine, meaning you only need to touch a few buttons to be rewarded with a well-made cuppa. It also means you can start the brewing process and walk away to complete other tasks, which is an excellent use of time if you ask me.
4400 Series Fully Automatic Espresso Machine: was $1199 now $613
Amazon has already discounted the 4400 to an accessible AU$763, but is offering an extra AU$150 off via a coupon, so don’t forget to check the box. And if you have a Prime membership, your new best friend can be up and running in your kitchen within 48 hours. View Deal
The Philips LatteGo 4400 Series isn’t the first automatic espresso machine I’ve used — and it certainly won’t be the last — but it’s left a lasting impression on me for a number of reasons. Firstly, it’s just so darn easy to master. Touch-enabled buttons are clearly labelled and the small colour screen is easy to read.
Then there’s the main event: the coffee. As I said in my Philips LatteGo 4400 Series review, it consistently brews a great-tasting coffee. I did have to adjust the grind setting for the best results, but thankfully that’s an easy process that shouldn’t scare newbies.
What you’re also getting here are 12 drink presets, two customisable user profiles and an automatic milk-foaming system, which is where it gets the LatteGo name. The user profiles are particularly useful if you have more than one person in your household using the machine: just enter a profile, select a drink and then the start button, that’s it.
Where I found the LatteGo 4400 falls a bit is with its headline feature — the LatteGo milk-foaming system. This sees a milk carafe being attached to the machine, for it to then deliver hot foamed milk into your cup or mug. I personally wasn’t all that thrilled with the level of foam produced, and it certainly won’t appeal to hardcore cappuccino fans, but flat white drinkers might actually prefer it. I found I achieved much better results using the Philips Baristina Milk Frother separately, and that’s not an expensive addition to add to your coffee routine, costing just under AU$120 on Amazon.
As for the actual coffee-making part though, I simply cannot fault the 4400 Series. Navigating through the menus is as easy as can be, and I’m consistently satisfied with the results.
While this isn’t the lowest-ever price I’ve seen on this highly capable machine, it comfortably beats the price I saw during Prime Day back in July (when it was AU$878). I can’t be sure how long this discount will last, so this could be one to jump on ASAP if your kitchen is crying out for a coffee machine.
Want more suggestions?Amazon has a few other coffee machines at discounted prices at the time of writing, including the Philips Baristina and Ninja Luxe Cafe, both of which claim spots in my guide to the best coffee machines in Australia.
Here are some other standout deals for you to consider:
- Philips Baristina – was AU$559 now AU$449.25
- Ninja Luxe Cafe (Stainless Steel) – was AU$799 now AU$619
- Breville Bambino Plus – was AU$729 now AU$498
- KitchenAid Semi Automatic – was AU$999 now AU$777
Demand for AI fluency has risen nearly sevenfold in two years, faster than demand for any other skill, according to the McKinsey Global Institute.
My conversations with technology leaders echo this: nearly every one I speak with is hiring for AI skills.
The problem, however, is that fewer can tell me what AI fluency actually looks like on the job at their organizations.
McKinsey found nearly 90% of companies have invested in AI. Fewer than 40%, however, report measurable gains.
Most AI post-mortems take a hard look at models, data and workflows.
Few look at the people hired to champion AI transformation or how and why they were chosen.
The risks of mistaking confidence for competenceA side effect of AI advancement is that it has made the ability to speak about AI use much easier. A candidate might start by naming every model, then walking you through an architecture they read about last week. In a short conversation, fluent language is almost impossible to separate from fluent practice.
Those who win over the hiring manager are those who sound the most at ease when speaking to AI. Whether they can actually use AI to do the job is a separate question, and most hiring processes never ask it.
Think of an interview as a demo on the candidate’s chosen grounds with controlled setup and their narration. On-the-job AI use is where the trouble starts: without proper AI fluency, it’s easy to lose control of messy data, missing edge cases, systems that refuse to talk to each other, or a model that hallucinates at the worst possible moment.
Just because a candidate dazzles in the demo doesn’t mean they won’t stall in a production environment.
On the surface, this looks like a people problem. In reality, it’s a hiring process problem.
What does a bad AI hire really cost?The true cost of bad AI hires typically takes time to manifest. It’s rarely on day one, and it can often take several forms. For example, a new AI model goes into production and starts hallucinating in front of a client.
Digging further into the root of this risk reveals that most companies have no shared definition of what “good” looks like for working with AI. One manager may test it one way, another may test it completely differently, and a third goes on gut feel. When the fluency bar changes with every interviewer, organizations are limited to collecting surface-level impressions rather than assessing skills.
Without that shared definition, the damage eventually reveals itself. A project stalls because the person who talked a great game can't get a model to produce anything reliable. Or worse: they can, but only for themselves. They 10x their output, then leave light documentation no one else can follow and route around security and compliance to do it.
Therein lies more risk: productivity that works for one person and breaks the system around them has wide-reaching ramifications.
Eventually, the work has to be redone and the role reopens. Project deliverables slip another quarter and the company is forced to burn more budget.
Bad hiring is not the only reason AI investment underdelivers, but it is a bigger one than most leaders admit.
Test for competence, not vocabularySimply adding another requirement to job descriptions isn’t going to solve this issue. Instead, it's deciding the scope of AI use required by the role, then building a process that makes job candidates demonstrate this level of use rather than describe it.
Too many companies set the standard around whether a candidate knows the tools exist. Going deeper to set the bar at independent, verified use makes the fluency gap immediately visible.
There are several changes teams can make to the hiring process to shift the measurement of AI fluency:
1. Test the work before the conversationPut candidates in front of a scenario that mirrors the real job and score it against a fixed set of capabilities agreed in advance. Say you’re hiring for a product designer, then drop them into a live brief with a real product constraint and ask "why" at every step.
The candidate limited to fluent talking hands you forty polished directions. The one who can actually do the job tells you which one survives the edge cases, what accessibility rules are needed, and the engineer who has to build it. That's the test most interviews never run.
2. Change the interview questionsStop asking which tools people use. Ask about the last time AI gave them a wrong answer and how they caught it, or to show a prompt that failed and how they fixed it.
Go deep on one real piece of work: what changed, what broke, what they checked, and who the output affected. None of that survives someone who has not done the work.
3. Pivot the scriptGive the candidate a realistic AI task, then change it halfway through by taking a tool away, adding a constraint, or moving the goal. Strong candidates reframe and carry on, while performers stall or describe what they would theoretically do.
Ask to see the workings, not the output. Let people use AI in the assessment on the condition that they show you their workflows, automation, prompts and the reasoning behind them. Anyone can produce a polished output now. How they got there is more important.
4. Score it togetherGive every interviewer the same dimensions to judge, set before the conversation, then compare evidence in a shared debrief session. The shared evidence review becomes the critical lever for the final hiring decision and fuels the definition of AI fluency for that candidate.
Today’s AI hiring continues to fall into the trap of selecting the candidate who interviews best, only to find a quarter later that they can’t deliver the actual work.
A vague hiring process may feel efficient because it asks less of the people running it. However, the business impact is real as it moves the cost downstream into deployment, where it becomes much harder to trace and far more expensive to fix.
Next time you make an AI hire, skip the questions about which tools they use. Instead, focus on the last time one of those tools failed them, and what they did about it.
We've featured the best HR software.
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