<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Tim Daines on Medium]]></title>
        <description><![CDATA[Stories by Tim Daines on Medium]]></description>
        <link>https://medium.com/@TimDaines?source=rss-322ceac276c7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*F1MXluVao-CTGb1Gt0_DlQ.jpeg</url>
            <title>Stories by Tim Daines on Medium</title>
            <link>https://medium.com/@TimDaines?source=rss-322ceac276c7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 21:30:32 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@TimDaines/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Is there such a thing as truly Green AI?]]></title>
            <link>https://medium.com/@TimDaines/is-there-such-a-thing-as-truly-green-ai-6deb2f7a24eb?source=rss-322ceac276c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/6deb2f7a24eb</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[climate-change]]></category>
            <category><![CDATA[innovation]]></category>
            <dc:creator><![CDATA[Tim Daines]]></dc:creator>
            <pubDate>Sat, 07 Oct 2023 09:23:01 GMT</pubDate>
            <atom:updated>2023-10-07T09:23:01.943Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2unxvLkBAos22Ygwe8bfdw.jpeg" /></figure><p>In every Annual Report, Jeff Bezos included his 1997 Letter to Shareholders. This letter emphasises customer focus, long-term value, and making bold bets. Bezos stated,</p><blockquote>This is Day 1 for the Internet and, potentially, for Amazon.com.</blockquote><p>With AWS’s introduction, Amazon’s security became pivotal to customer trust, marking Security as Day 0. Considering the rising energy demands of advanced AI like large language models, are we now seeing a new day emerge for AI product teams — GreenAI as Day -1?</p><h3>The Growing Carbon Footprint of AI Models</h3><p>Today’s cutting-edge AI models have an enormous appetite for computing power, but this comes with a huge cost to carbon. Let’s take GPT-3 and the training of this large language model (LLM). It is estimated the <a href="https://www.forbes.com/sites/robtoews/2020/06/17/deep-learnings-climate-change-problem/?sh=2f6039766b43">emissions</a> generated from 175 billion parameters, consumes approximately 1,287 megawatt hours of electricity and generate 552 tons of carbon dioxide. This is the equivalent of 123 gasoline-powered passenger vehicles driven for one year.</p><p>These massive models devour electricity through their training computational requirements. Leading models are trained on thousands of GPUs running nonstop for weeks or months.</p><p>For example, <a href="https://medium.com/codex/gpt-4-will-be-500x-smaller-than-people-think-here-is-why-3556816f8ff2#:~:text=Training%20100%20Trillion%20Parameters,GPUs%2C%20would%20take%2053%20Years.">training GPT-3</a> took 1024 GPUs, took 34 days, and cost $4.6M in compute alone. Training a 100 trillion parameter model on the same data, using 10000 GPUs, would take 53 years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/815/1*ZBTO2gtPYJ_17F_nOUNXNQ.jpeg" /><figcaption>Source: <a href="https://cirrascale.com/cerebras.php">https://cirrascale.com/cerebras.php</a></figcaption></figure><p>To put that into perspective, training a single large AI model can use more power and is <a href="https://www.techtarget.com/searchenterpriseai/feature/Energy-consumption-of-AI-poses-environmental-problems">five times</a> the lifetime emissions of an average American Car (including manufacturing of the car itself).</p><blockquote>The largest models being developed today are in the orders of magnitude more complex than GPT-3.</blockquote><p>As AI capabilities advance, energy and emissions scale up. The computational complexity underlying models like GPT-3 means AI systems could end up consuming over 5% of the <a href="https://www.bloomberg.com/news/articles/2023-03-09/how-much-energy-do-ai-and-chatgpt-use-no-one-knows-for-sure?leadSource=uverify%20wall">world’s</a> electricity within a decade.</p><p>We’ve entered the era of the gigantic carbon footprint model, but what practical steps can we put in place today to curb these emissions? The first step is acknowledging the invisible environmental impacts of developing ever-larger AI systems.</p><h3>The Overlooked Impacts of AI Innovation</h3><p>The race to develop ever-more advanced AI capabilities has come at a cost:</p><blockquote>Sustainability, an afterthought.</blockquote><p>Businesses are in the pursuit of unprecedented model performance, but their approach creates issues surrounding energy efficiency, emissions, and environmental impact have been largely ignored. This prevailing mindset to throw more data and computing power at AI to achieve new breakthroughs in accuracy and capability is a single-minded focus on pushing the limits of what’s possible, but has made AI innovation akin to a runaway train, barrelling ahead without pausing to consider externalities.</p><p>Most development teams don’t <a href="https://greensoftware.foundation/articles/uniting-for-the-green-software-revolution">track</a> the computational resources or emissions associated with their models. These environmental impacts remain invisible, allowing AI’s energy appetite to balloon unconstrained.</p><p>What’s desperately needed is a widening of perspective in the field. If AI is to fulfil its promise responsibly, sustainability must be embedded in the development process from day -1, as opposed to an add-on or box-ticking exercise.</p><p>Rather than overlooking sustainability, business leaders must make it a core consideration and design constraint. This holistic systems view will enable the emergence of AI that drives progress through efficiency rather than brute compute force. The potential is there. But reframing innovation in AI will require a collective commitment to lift our heads up from narrow accuracy gains when considering technology’s complex environmental interdependencies. If we meet this challenge, simply AI today will become green AI tomorrow.</p><h3>Making AI More Sustainable, Starting Today</h3><p>While the environmental impacts of unrestrained AI progress seem daunting, there are practical steps AI product teams can take today to steer innovation in a greener direction:</p><ul><li><strong>Prioritise energy efficiency in computing</strong> — choose cloud vendors who optimise for performance per watt, and who retrofit existing systems with power-saving chips and software.</li><li><strong>Build smaller, lighter models</strong> — developing techniques like knowledge distillation to slim down models without losing accuracy is a good start. Using smaller datasets and model sizes from existing already curated reduces carbon and speeds up experimentation.</li><li><strong>Adopt carbon-aware development practices </strong>— in your AI product lifecycle process, ensure emissions are tracked throughout the lifecycle. Set go/no go carbon metrics for your project.</li><li><strong>Use efficiency-focused frameworks</strong> — Edge device technology is hugely benefiting from Ai today. Both <a href="https://www.tensorflow.org/lite">Tensorflow Lite</a> and <a href="https://pytorch.org/mobile/home/">Pytorch Mobile</a> enable optimising model performance, reducing latency, preserving privacy, and enable low carbon interactive use cases.</li><li><strong>Shift compute workloads to renewable energy sources</strong> — training models take time and heavy compute. Times and locations of where renewable energy is abundant hugely benefit reductions in carbon usage. If data sovereignty locations are not an issue, consider this setup when training models.</li><li><strong>Open source systems and publish energy usage data</strong> — best practice sharing of sustainability will be the key to getting to green ai. The starting point is to increase transparency.</li><li><strong>Join green AI initiatives</strong> — <a href="https://greensoftware.foundation/">The Green Software Foundation</a> is leading the way to exchanging best practices to reduce AI’s carbon footprint.</li></ul><p>While daunting, climate concerns also represent an opportunity to rethink how AI is built. With a renewed focus on efficiency and sustainability, the field can advance in more responsible ways. But it starts with each AI product team taking an honest look at their own carbon impacts, and taking steps today to lighten them.</p><h3>Moving Toward Responsible, Eco-Friendly AI</h3><p>While optimising today’s systems is important, truly green AI products require starting from scratch.</p><blockquote>What’s needed is a cradle-to-grave approach that bakes in sustainability from day -1.</blockquote><p>This begins at the discovery phase, evaluating the full lifecycle emissions of different model architectures and hardware configurations. It also means holistically tracking emissions at each stage of the AI product lifecycle — from data collection through model training to user inference in production.</p><p>At every part of the AI product process, the product team should default to the most efficient methods using the minimal computing resources feasible. Data pipelines should be engineered for energy conservation, not just speed and throughput.</p><p>Sustainability thinking also requires questioning the scale of models and datasets as they grow with efficiency and accuracy.</p><ul><li>Do they need to be so large?</li><li>What is the carbon cost of marginal accuracy gains or model capacity?</li><li>Would a smaller model suffice?</li></ul><p>Finally, sustainability should be monitored just like any other product, OKR. Teams can set carbon budgets for projects, and treat emissions as a key metric to optimise.</p><p>This product lifecycle view reaches beyond design, development and into deployment. Features like user carbon transparency and low power mode can further optimise production systems.</p><p>The goal of any AI product team should be about green considerations intrinsic to how AI is conceptualised and built. By embedding sustainability in from the start, AI can advance responsibly and minimise unintended impacts.</p><h3>Are You Ready To Go Greenai?</h3><p>The path toward sustainable AI will require collective effort across all spans and layers of a business, as well as external communities. While the challenges seem daunting, each of us can make a difference through our everyday choices and development practices of AI.</p><p>Here’s how you could get started:</p><ul><li>First, educate yourself on the carbon impacts of AI and strategies for green technology. Groups like the Green Software Foundation provide courses and resources to get started.</li><li>Second, advocate for sustainability in your team and organisation’s AI product roadmap. Raise awareness, spark discussion, and propose ideas to reduce emissions.</li><li>Third, adopt green tactics in your own work such as optimising for efficiency, using smaller models, and shifting intensive workloads to renewable energy sources.</li><li>And finally, support public policies and industry standards that promote transparency and accountability around AI’s environmental footprint.</li></ul><p>By taking small steps today, we lay the groundwork for more responsible innovation tomorrow. The future of AI is not predetermined — it will be shaped by the priorities and values of people like you.</p><p>Let’s build a green brick road.</p><p>Here’s how I’m helping companies go GreenAI — <a href="https://www.zombiemolecule.com/workshops">https://www.zombiemolecule.com/workshops</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6deb2f7a24eb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The case for service design when designing connected digital health systems]]></title>
            <link>https://medium.com/@TimDaines/the-case-for-service-design-when-designing-connected-digital-health-systems-8c178834f470?source=rss-322ceac276c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/8c178834f470</guid>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[interaction-design]]></category>
            <dc:creator><![CDATA[Tim Daines]]></dc:creator>
            <pubDate>Tue, 08 May 2018 15:49:22 GMT</pubDate>
            <atom:updated>2018-05-08T22:32:10.767Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FBSJZzgIRMFPnPkHwjlK3g@2x.jpeg" /></figure><p>The patient pathway is becoming increasingly digital and with the number and types of digital touch-points ever increasing, vast volumes of data are being generated opening the floodgates to new and emerging technologies that provide patients and healthcare stakeholders, with better access to information when managing health.</p><p>Depending on the moment-in-time, a patient can now interact around a single touch-point (perhaps a medical device with sensors on) or across a multiple set of touch-points, but for the desired patient experience to add-value all of these touch-points must work together. Let’s take an example of a connected inhaler project I worked on where an asthmatic with 20% of their drug remaining in their inhaler must begin to think about a represcription. In order for the patient to receive a new inhaler on time and avoid affecting their dose regime, the connected digital health system must issue a repeat prescription within an acceptable time limit. The trigger to this chain of events starts with a metered level dose sensor where it sends data to an app when 20% remains, the app then notifies the batch system (via the cloud) informing it of the remaining 20%, a new drug prescription request is triggered and must be approved by the payer and clinician, upon approval the pharmacist receives a request to prepare a new prescription, where the patient then receives confirmation of the prescription either via collection or delivery to a location.</p><p>This system example is one of many I’ve been designing with colleagues over the last two years. But whilst such a system may promise opportunities to help patients and other healthcare stakeholders make better decisions, live and work more effectively in their job roles, and improve partnership operations, significant complexities designing for such touch-point ecologies quickly became apparent. Service design was at the summit when navigating these complexities, driving design thinking well before deploying both physical and digital medical devices into the hands of end users – and here’s why.</p><h4>Seamless Experiences</h4><p>Let’s begin with designing for a seamless experience. For a patient, designing in silos and then bringing touch-points together afterwards simply won’t work – from first hand experience this fails quickly and impacts not only design and development costs, but also the end-user value proposition. From the outset co-creation with the patient and stakeholders who affect the patient pathway must be included in order to discuss problems together. But such an approach at first invites huge resistance, namely from HIPPOs (Highest paid earners) in the room. It took the communication of three failed digital health service case studies to get the buy-in of executive leaders for our inhaler project, where we could then hear personal health stories of those who work in and try to manage asthma. The value of these stories alone brings the realisation of what kind of service needs to exist and who will eventually own, scale and run it when it goes live.</p><h4>Design Approach</h4><p>Then there’s the design approach. I’ve used waterfall, agile and lean approaches across a multitude of projects, but as a service designer I was often conflicted in conversation with industrial designers, engineers, developers and UX designers. It was a challenge to understand when to use waterfall, lean and agile all together, whilst making sure everyone’s voice was heard, as well as understanding the different design language being batted in design reviews. Physical sensors, smartphones and cloud-based internet services forming communities of knowledge, also challenge the design approach because they all do something with data.</p><blockquote>Never underestimate the movement of data autonomously or synchronously around the real-time use of a connected digital health system.</blockquote><p>Whilst the data itself may well be rich in insights where its collected and used to offer guidance, and accurately diagnosing around certain points of the patient pathway to solve various problems, I found myself asking questions around the intent patients and healthcare stakeholders may have with the data, and why it is important for them to engage with it. Furthermore, with my interaction design hat on, there is a role here in supporting the development, delivery and experience across the service when in a ‘live’ state of silent data intelligence and when the experience is invisible to end-users.</p><h4>ML, AI and Blockchain</h4><p>Of course, lets not forget the currently hyped Machine Learning, Artificial Intelligence and Blockchain. These have been a vital ingredient into digital touch-points when running parts of a connected digital healthcare system. But whilst this promise of opportunity through these technologies may help patients and healthcare stakeholders ask questions they never thought of, make decisions and become more efficient over their careers and their lives, there are challengs in understanding how to design the experience based on how various data is being used in the algorithms. Using existing mobile design patterns quickly failed when algorithms were dictating the movement of information in our inhaler project. It wasn’t until a new set of best design practices through ‘service safaris’ from ‘in-the-wild’ observations and roleplay began to form, where we ended up leaning towards behavioural economics and jobs-to-be-done.</p><p>Designing connected digital health systems that use sensors, the web, ML, AI and Blockchain to form touch-point ecologies will require everything to be connected, as everything needs to think and act as one whole system. The Internet of Things is enabling much of this change, but service design thinking whether as a mindset or in the form of experienced professionals is the ultimate summit of designing these systems, as both interconnections and relationships must co-exist.</p><p>What do you think?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8c178834f470" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why there’s no ‘silver bullet’ to creating a connected drug delivery system]]></title>
            <link>https://medium.com/@TimDaines/why-theres-no-silver-bullet-to-creating-a-connected-drug-delivery-system-df5e58087de4?source=rss-322ceac276c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/df5e58087de4</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[connected-devices]]></category>
            <category><![CDATA[iot]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Tim Daines]]></dc:creator>
            <pubDate>Tue, 27 Mar 2018 12:49:19 GMT</pubDate>
            <atom:updated>2018-03-27T14:08:13.716Z</atom:updated>
            <content:encoded><![CDATA[<h4>By Tim Daines, UX and Service Designer – Digital Health</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/303/1*5R01leIsV43JHi8TMS_9Ug@2x.jpeg" /></figure><p>If you hadn’t noticed, drug delivery devices are going through a disruptive shift where regulation, partnerships with healthcare players, tech giants and startups, are all foraging in the space of AI (Artificial Intelligence), Blockchain and IoT (Internet of Things). This shift is creating a strange new world, one where the drug delivery is now smart enough to analyse the biological properties of a drug - once entering the body, is close to expiry, or even spoilt by temperature.</p><p>The drug delivery device is no longer a solo product, it is a companion to a smartphone or smartwatch, and part of a network sending data to manifest many interactions – direct with patient, others in healthcare, and with the payer – across many contexts, interfaces and interactions. The UX challenge for Industrial, GUI (Graphical User Interface) and VUI (Voice User Interface) designers can no longer be designed in a silo, and must be combined from the conception of defining the problem in behavioural terms. The permutations with such complexity for those designers who consider themselves the very best with their individual craft, will result in unintended consequences across integration, behaviour economics and human-centred design.</p><p>No silver bullet exists to create a fully integrated drug delivery device, and whilst industrial and digital designers may have good intent for users, what all must realise is that they are creating a connected system and not an integrated drug delivery product.</p><p>As a UX and Service designer of connected systems, here is part one of a two part piece, focusing on the mistakes that could result in a drug delivery system failing users, employees, commercial operations and technology:</p><ul><li>Mistake #1 – Creating a gap between human and machine</li><li>Mistake #2 – AI will save us all</li><li>Mistake #3 – Acquiring data before knowing the problem its trying to solve</li><li>Mistake #4 – Not designing for death</li><li>Mistake #5 – IoT is just another feature</li></ul><h4>Mistake #1 – Creating a gap between human and machine</h4><p>Ask yourself this ‘What does Patient engagement mean to you?’ Our planet is in a state of flux when it comes to healthcare today, and the demands on our existing systems are too great to ignore. Whilst the heralding of being able to make healthcare more intelligent through AI (when properly matched to a problem) and optimal through IoT into workflow, the broad scope of mobile health (mhealth), health information technology (IT), wearables devices, telehealth, telemedicine, and personalised medicine, is establishing an ever increasing divide between human and machine.</p><p>This is not uncommon. Edutech, ecommerce, energy, banking, to name but a few, have all fallen on this sword. Sure enough, digital has the potential to transform how patients and carers can consume health information to act on, but a connected drug delivery system will not be an integrated physical artifact where a single transaction occurs. It will exist within a ‘lived’ context and will be more intimately associated with the body and a patient’s personal data. Patients will need to be equipped with a seamless user experience that provides confidence and capability to access information, where empowerment in self-management to achieve greater well-being will be the measure of success – after all the <a href="https://www.amazon.co.uk/Patient-CEO-Technology-Empowers-Healthcare/dp/161961376X">patient is the CEO</a> of their own body, not the designer telling them what they need from their own experience.</p><p>If you do not understand how a patient and carer currently manage their disease and go straight to the ‘how’ and ‘what, you will affect business and technology factors that devalue the proposition behind the intent of the drug delivery system. A real-world example of this is Propeller Health’s connected inhaler system, which had seamless issues when rechargeable batteries and manual bluetooth connecting were introduced when connecting with a <a href="http://www.mobihealthnews.com/content/three-lessons-propeller-learned-decade-digital-health">smartphone</a>.</p><h4>Mistake #2 – AI will save us all</h4><p>There is already enough bumbling going on out there with the huge potential AI can bring to healthcare. Will it really replace doctors? What will care post usage using AI look like? But is ‘Artificial Empathy’ being thought about during the creation of a drug delivery connected system? If AI is to deliver meaningful and actionable insight where all doctors and carers know the affects of their therapy, underlying questions need to be asked to determine how AI will meet the personal ‘emotional’ needs of every stakeholder in the healthcare system.</p><p>While AI could have the potential to provide significant opportunity in the connected drug delivery space, two challenges in present day culture still impact this potential: Dealing with integration into a healthcare ecosystem currently stuck in the ‘stone age’ due to its reliance on legacy technology (fax machines and pagers); closing the ‘empathy gap’ between AI and human before it’s benefits are realised. Tim Caynes and Jane Vance provide an in-depth view on this latre point when designing AI in healthcare solutions – ‘<a href="http://www.foolproof.co.uk/thinking/the-empathy-gap-designing-intelligent-healthcare-systems/">empathy gap</a>s’ are more apparent than we think.</p><h4>Mistake #3 – Acquiring data before knowing the problem its trying to solve</h4><p>I always reflect back to three questions my mentor drummed into me before embarking on any design work:</p><ul><li>What problem are we trying to solve?</li><li>How do we know this is the problem?</li><li>When will we know we have solved the problem?</li></ul><p>I am not a life scientist, but the social scientist and designer in me suggests it should be obvious that any sensor being functionally designed into a physical drug delivery device to acquire data, has to be evidence-based in line with medicine-based counterparts. We’ve seen those headlines ‘<a href="http://dataconomy.com/2017/08/5-ways-healthcare-big-data/">Big Data will save billions through health analytics</a>’, but whilst real-world evidence through data-driven design could improve efficiency, disrupt clinical trials, and get digital health products faster to market, how will data add value to patient and carer to solve problems with disease management?</p><p>Let us imagine we have an inhaler functionally designed with a precision based gyroscope sensor in it, enabling data collection to measure whether inhaler use technique is adequate enough by the patient. During it’s use, data is diligently being collected and cleaned where well-founded data science is being applied to a resulting dataset. But what if the analysis of this data determines the evidence to be inconclusive? What does the patient and even the medical specialist do at this point? If we want to achieve a paradigm shift where the data is meaningful to solving problems, we need to first understand why, how and what data is going to be needed. Just one dataset from our gyroscope example is not going to be enough to determine whether a patient has correctly ingested their drug.</p><p>Data is the glue that binds a connected drug delivery system together, and if its understood why it is needed then the right combination of data collection mechanisms can be identified i.e. functionally designing in the right precision based sensor. Data analysis also needs to be managed, as it will be agile and evolutionary through its iterative collection. If management of data is not considered during the design of the connected drug delivery system, new challenges will appear in employee operations where a ‘reactive’ rather than ‘proactive’ approach will occur.</p><h4>Mistake #4 – Not designing for death</h4><p>As morbid as it sounds, death will happen to all of us. It is a simple end-to-end design really – you are born, you live, and then you die. In our current industrial system, during our living stage, we are all expected to write a ‘Will’, ensuring our estate of physical assets is left to the ones we love. Now in a world of digital, our lives no longer begin and end in a physical place, as we have established digital footprints made up of pixels and data around us that are still prevalent once we have passed.</p><p>Facebook profiles are a prime example of where we establish and leave our digital footprints. I recall a friend several years ago who sadly died in a tragic car accident, and like most would think saying farewell at a funeral brought closure to our friendship. We were friends on Facebook though, and I am often reminded of there existence because the algorithms driving the site only know when we have not spoken over a long period of time – ‘It’s been a while, reconnect with Mike.’ When I am reminded in this way my first thought is great sadness of how short their life was, but my second thought is ‘how do I stop this without unfriending them and losing visibility of the memories we had together?’</p><p>If these design patterns continue into connected drug delivery systems, we will surely hit the same problems Facebook experience – stagnant user profiles. For example, in a non-connected state a patient who has passed away from COPD (Chronic Obstructive Pulmonary Disease), will require the carer(s) to notify organisations to cease payments, benefits and taxes. In a connected state this will go beyond the bureaucracy of benefits and taxes – who owns the data assets a patient has? How does a patient decide who gets their data? What happens to the data when that patient eventually passes?</p><p>Connected drug delivery systems are rapidly becoming their own ecosystems, and whilst solutions may have been implemented with various value propositions to get them used, their design may suffer as a result of algorithms failing to recognise a patient has passed away.</p><h4>Mistake #5 – Adding IoT as another feature</h4><p>As tempting as it might sound, adding a physical sensor to connect an existing drug delivery product to the Internet is not the shift we think IoT brings. If the user experience has been put at the centre of a connected drug delivery system, what we should see is physical controls being designed in conjunction with its digital component. This way we provide a seamless GUI of consistency where terminology, visual language and interaction patterns, will not disrupt the flow of the user experience.</p><blockquote>IoT designed into a drug delivery system is about providing an intervention that could not have previously been achieved in the management of a disease. It must be invisible to the point where it is inclusive and not intrusive. – Tim Daines 2018</blockquote><p>IoT is rapidly evolving into an ecosystem where new business use cases are emerging to improve stakeholder efficiencies around the management of a disease. If physical sensors are added to a physical drug delivery device in isolation of the connected system, additional consciousness on top of existing disease management will be introduced where the likelihood of failure of the system will be significantly high.</p><p>People designing connected drug delivery systems in healthcare who want to provide frictionless experiences for patients and stakeholders are being worked on – but before these systems are designed, the digital health story must first be understood. If patients and their carers health stories are ignored from the outset, the true value of a connected drug delivery product or service may never be realised.</p><p>Part two will discuss issues around scaling, organisational design challenges, GDPR, and how we should not be complacent with online infrastructure!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=df5e58087de4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You talkin’ to me?]]></title>
            <link>https://medium.com/@TimDaines/you-talkin-to-me-57d7c64178da?source=rss-322ceac276c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/57d7c64178da</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[amazon]]></category>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[vui]]></category>
            <dc:creator><![CDATA[Tim Daines]]></dc:creator>
            <pubDate>Tue, 03 Jan 2017 14:41:14 GMT</pubDate>
            <atom:updated>2017-01-03T14:41:14.469Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*XHLR1x29-bGg7ZCbaCcEXQ.jpeg" /><figcaption>Labelled for non-commercial reuse — <a href="https://pixabay.com/en/photos/speaking/">https://pixabay.com/en/photos/speaking/</a></figcaption></figure><p><strong>7 principles when designing for voice user interfaces (VUI)</strong></p><p><a href="http://www.amazon.co.uk/Amazon-SK705DI-Echo-Black/dp/B01GAGVIE4">Amazon’s Echo</a> and <a href="http://www.apple.com/uk/ios/siri/">Apple’s Siri </a>have blazed the trail. Voice input and control have reached the technology mainstream and a new generation of digital services and physical devices will soon be upon us. But for those of us in technology development, how should we go about designing an experience that knows us as well as our mother?</p><p>There’s much discussion around the subject of voice-user interface, or VUI for the acronym lovers; but nobody quite knows what’s in store for the future. One thing we can be sure of though is that it’s not the shiny ‘hard’ design that matters here. Instead it’s the ‘soft’ design which must be perfected and made delightful; specifically it’s a ‘service’ that uses ‘audio data’ to assist someone in their life. This ‘service’ will provide a user experience, but equally it’s the ‘audio data’ design challenge which must be worked through, particularly as there aren’t any screens that a user can interact with.</p><p>These are early days for VUI, and best practice is yet to be established. With that in mind, I wanted to propose my own core principles when designing for voice-user interface. Some of this thinking is the result of having recently attended a user experience hackathon, focused on designing for VUI.</p><p><strong>1. Don’t build a service like Fawlty Towers</strong></p><p>The legendary UK TV series <a href="http://en.wikipedia.org/wiki/Fawlty_Towers">Fawlty Towers</a> follows the exploits and misadventures of a short-fused hotel owner named Basil Fawlty. His relationship with the waiter Manuel is comically confused and irritated. Basil’s frustration with Manuel grows based on constant complications and mistakes in ‘dialogue’ due to cultural language barriers (he’s “<a href="http://www.youtube.com/watch?v=0kqyev46qyI">from Barcelona</a>”). The result is escalating frustration and a one-way shouting match. Understanding how to design for dialogue will be key to ensuring your VUI returns a response it’s confident in and which supports the user’s needs. By making <a href="http://en.wikipedia.org/wiki/Dialogue">dialogue</a> a core design principle, technology developers have a chance of building products that users can enjoy interacting with. Beware though, <a href="http://www.youtube.com/watch?v=n0dz1GTzRTk">even the very best get this wrong</a>.</p><p><strong>2. Look mum, no hands!</strong></p><p>Remember we’re designing a VUI product that will listen to the users in a variety of conditions. There’s no machine equivalent of the <a href="http://en.wikipedia.org/wiki/Cocktail_party_effect">cocktail party effect</a>, whereby we humans can focus our auditory attention on a particular stimulus while filtering out a range of other stimuli. VUI design will need to be defined around how normal human behaviour tunes into a single voice. Understanding this dialogue experience is vitally important, as it will inform the ‘hard’ design in relation to audio source separation, including the combination of intelligent signal processing and the placement of microphones. Without understanding how noise suppression will be designed to immediately detect directed user dialogue, it’s likely that the entire ‘soft’ design will fail.</p><p><strong>3. It’s the 21st Century</strong></p><p>Have you ever wondered why Amazon Echo defaults to a female voice? Is it because females are more helpful and males are not? Take great care when choosing the voice that speaks to the user. We are a complicated species with a mix of cultures, languages and preferences. The choice of default VUI voice could easily send a message we didn’t intend.</p><p><strong>4. Why complement the VUI with an App?</strong></p><p>I have serious reservations about the value that <a href="http://itunes.apple.com/gb/app/amazon-alexa/id944011620?mt=8">Amazon’s Alexa app</a> for Echo brings to the user. It’s useful during setup and acts as a repository for past question and answers, but Amazon are introducing further complexity into an already unusual and complex system, and designers run the risk of cognitively overloading the experience of the VUI. Without wishing to be too purist about it, I’d be happier if VUI products and services were so unique as to be impossible to replicate with a simple app.</p><p><strong>5. The user’s journey is VUI not GUI</strong></p><p>In the world of graphical-interface design (GUI) I’ve come to understand a rule of thumb that suggests our work is 80% design and 20% experimenting with the user. But with voice-interface design we’re not designing and experimenting with interface design, rather we’re designing for dialogue. My own experience is that this completely upends the old rule of thumb, moving us to 80% experimenting and 20% design. Remember, we’re designing for what the user is saying (conversational and transactional), and not what the user is doing (interaction and navigation). This shift to a focus on experimentation and learning then leads to a complete renegotiation of the relationship between designer and developer across the entire product strategy. Don’t take this lightly.</p><p><strong>6. Build for empathy</strong></p><p>Whilst a VUI system won’t (yet) know how the user is feeling, the dialogue and perceived character of the VUI will be important when responding to the tone of voice it understands from the user. This understanding should have been learned and improved over time from numerous conversations that take place between the user and the VUI. In short, the conversation needs to be as natural as possible; humans intelligently match their tone of voice to those they’re speaking with as a way of putting people at ease. The best machines will do the same — this is part of what the vital Machine Learning element of such services can deliver.</p><p><strong>7. VUI needs to mature with the user</strong></p><p>Cloud computing has matured and is all-pervasive. Users are now able to take their data with them when they upgrade their ‘hard’ design product to the latest technology. This must be true of a VUI technology as well, particularly when concerning the preservation of numerous conversations. Whilst this ‘hard’ design element — the physical thing — may remain locked in conventional product lifecycles, the ‘soft’ design part — the digital service — should have loftier ambitions and a longer lifespan. It should evolve its existence based on learning how the user’s dialogue changes as they themselves evolve and age. Truly visionary VUI will design for this process of adapting, learning and understanding the biological evolution of its user.</p><p>An area I haven’t examined in detail here is AI, and the ways in which Machine Learning and the right inputs and outputs might enable VUIs to replicate the way a human mind thinks. Machine Learning today is improving VUI, but not convincing anyone that they’re speaking to a real human. Indeed it may be that <a href="http://blog.cambridgeconsultants.com/wireless-product-development/turing-services-is-that-a-human-or-an-algo/">a real human hand is delivering some VUI improvements</a>.</p><p>In addition, as VUIs become better at detecting the emotions of humans we must be mindful of the ethical considerations at play. Should we allow VUIs into the medical space, for instance, allowing them to give advice on medical conditions such as depression, perhaps even making referrals or advising a next of kin? It’s a fair bet that voice tools aided by Machine Learning will eventually be able to infer a long term deterioration of the user’s mood, so what should be done with such knowledge?</p><p>These principles are an early attempt to codify the core design principles when designing for voice-user interface (VUI). They’re not comprehensive and they’ll evolve over time. In the meantime, share your thoughts and any that I might have missed.</p><p>The original blog post can be found <a href="http://blog.cambridgeconsultants.com/wireless-product-development/you-talkin-to-me-7-principles-when-designing-for-voice-user-interfaces-vui/">here </a>at Cambridge Consultants.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=57d7c64178da" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Brands Big Data Crisis: The broken promise of anonymisation to create personalised services.]]></title>
            <link>https://medium.com/@TimDaines/brands-big-data-crisis-the-broken-promise-of-anonymisation-to-create-personalised-services-4a8b7092559b?source=rss-322ceac276c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/4a8b7092559b</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[leadership]]></category>
            <dc:creator><![CDATA[Tim Daines]]></dc:creator>
            <pubDate>Tue, 29 Mar 2016 22:09:48 GMT</pubDate>
            <atom:updated>2016-03-29T22:09:48.776Z</atom:updated>
            <content:encoded><![CDATA[<p>Over the last twelve years I’ve worked closely with learning and content management systems to solve many interesting and complex problems for well known global brands and some startups. I‘ve developed quite the heuristic process:</p><ul><li>I find out the business and user needs</li><li>I take a look at new or existing system information architecture (get amoungst the detail)</li><li>I work out what can be added and accessed in that system</li><li>I create a content map and strategy (Find an excuse to use LEGO to label my inventory)</li><li>I design the experience.</li></ul><p><em>The reward</em> - working with amazing people to build, measure and iterate something of significant value to a consumer.</p><p>Recently, I’ve been reflecting on a number of projects I’ve worked on. As I was waded through the many systems I had encountered, I began to see a pattern emerge. When I analysed this pattern it resembled a process I’ve become familiar with when working in startups —pivot and scale. What I had seen were datasets of consumer data <em>pivot</em> in physical local based databases, and now in the smart technology age <em>scale</em> to cloud-based data services.</p><p>Strangely, I did not feel the result of this <em>scale</em> to be an evolutionary milestone of the human species. What I actually felt was a crisis looming for brands and governments.</p><h4><strong><em>Establishing consumer trust</em></strong></h4><p>Brands want consumers to give them their personal data so they can package it up and sell something of value for even more personal data. Whilst this is a very simple business model, it has reached saturation and started to show signs of instability for brands:</p><ul><li><strong>Consumers continue to share more and more of their personal data more than ever: </strong>The cloud has established a new eco-system and directly influenced and modified consumer behaviour. For brands, this has enabled them to collect data through numerous omni-channels at the expense of being overloaded with more personal data than they know how to manage. Much like global climate, brands have lost control of what disruption they are causing around consumer privacy and security.</li><li><strong>The Anonymisation Lie</strong>: As brands continue to develop incredibly detailed rich datasets spanning multiple-layers in the cloud, a new science of de-annoymisation has rendered annoyminity meaningless. Paul Ohm provides a great paper on this <a href="http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1450006">new emerging problem</a> in the commercial sector.</li><li><strong>Smoke and Mirrors around Trust:</strong> There is sentiment by a consumer when they tick that terms and conditions box to hand over their personal data, as they feel secure by todays public awareness surrounding big data security and privacy. However, brands are losing sight of establishing trust with consumers, as the consumer is lost with what kind of personalised service they are actually receiving.</li></ul><p>These are new challenges that brands face, and consumers will begin to see contradiction if brands cannot conduct themselves professionally by communicating this correctly - A broken promise of anonymisation of big data has begun and is creating a rift of trust.</p><p>Solutions need to be found and culturally addressed if a brand wants to avoid this looking crisis in the big data world, in particular:</p><ul><li>Where are malevolent things happening within a brand?</li><li>How is personal data being protected across a host of omni-channel devices, whatever their make, model and age are?</li><li>Do brands fully understand the eco-systems they are developing as they transition from analogue to digital business strategies, where they can confidently answer <strong>WHY</strong>, <strong>HOW</strong> and <strong>WHAT</strong> they mean by a personalised service for their consumers?</li></ul><p>By collaborating with consumers outright on how cloud-based data services will use personal data, a partnership can be established whereby trust can be built. Spotify and Netflix have begun to leverage this around the Smarthome of the future, by figuring out human habits and providing products and services that bring value in the form of Mixed Tapes (Social Playlists) and Film suggestions (We think you’ll also like).</p><h4>Cloud-Based Dataset Design Principals</h4><p>Like all good design, design principles must exist to stay in scope of a brand’s goals and the consumer needs. This is no different when working with data, and if brands want to be seen as conducting themselves in a professional manor when using consumer data they could consider some of the following principles:</p><ul><li><strong>The quality of the data needs to be sound </strong>— It doesn’t matter whether you have the top developer talent in the world who will code that kick arse algorithm to power that App you must provide users. If the data doesn’t exist across all the layers to personalise your service to a consumer — it will fail.</li><li><strong>Bias and Variance of data must not exist</strong> — An engineers worst nightmare to the point where they are literally sweating and being asked to create bias in data, as holes either exist across multiple layers of datasets or the make up of dataset in one layer is weak. It may feel like the right thing to do when time isn’t a luxury and you just have to get it out the door, but at what consequence? The end result could have huge implications on the trust you are trying to build, where your consumer will come once and never return.</li><li><strong>Security and Privacy are a human right</strong> — An incredible amount of data is being handled across multiple layers, and much of it is being collected directly and indirectly from consumers. Anonymisation of data may be happening and whilst your average consumer may not care what is being done with their data, its a forgone conclusion that the triangulation of consumer data in cloud-based data services will de-anonymise their identity. For example, <a href="http://lifehacker.com/how-retail-stores-track-you-using-your-smartphone-and-827512308">emotional engineering</a> of data is a new approach being tested in shop stores that identifies a consumer through their face and mobile device.</li></ul><p>Personalising a solution for a consumer needs a high degree of thinking around the actual data points required, and the fewer the data points to determine the right patterns the easier a personalised solution is going to be. Brands will get this wrong by the way, and they’ll need to work out how to deal with it. I’ve found these approaches have helped to avoid these problems:</p><ul><li>Deal with the complexity of the data by creating an <a href="http://www.uxbooth.com/articles/complete-beginners-guide-to-information-architecture/">IA map</a> for internal stakeholders to have those business requirement conversations.</li><li>Create a <a href="http://uxpamagazine.org/creating-a-dogo-map/">DoGo</a> map to work out all of the tasks a consumer will be need to do, not only to keep your product and service simple but to also indicate how data will be collected and used.</li></ul><h4><strong><em>Communicating, Collaborating and Measuring the success of cloud-based data services.</em></strong></h4><p>When it comes down to data and the sharing and collecting of it in cloud-based data services, the most important thing for brands to do is to start simple. This is key for any brand regardless of where they are as a business — startup or commercial. If brands are serious about establishing trust with their consumers, they need to step back and do the research before leaping in and doing everything quickly. The days of “Let’s get something done and up in-front of consumers” is pro-longing the long term damage of a brand in the digital landscape.</p><p>Over my twelve years working with data in learning and content management systems, I’ve come to find that those brands successfully connecting with their consumers have taken the following approach:</p><ul><li>Build a Plan,</li><li>Understand the data,</li><li>Understand the current company culture,</li><li>Understand the current consumer culture.</li></ul><p>Once all of the research has been done at this stage, then comes the internal buy-in and that sponsorship from a C-suite level leader — It’s key to do this if a brand wants to survive long term. When the new infrastructure begins to churn the re-ordered cogs and is integrated at a pilot level, get it tested, get it measure, learn from it and iterate it.</p><p><strong>What to measure?</strong></p><p><em>Purpose</em></p><p><a href="http://www.apple.com/uk/researchkit/">Apple’s ResearchKit</a> has introduced a brilliant ethos when concerning the use of consumer personal data. By being upfront with consumers what ResearchKit does - talking about security and privacy as core data design principles - this has enabled consumers to develop trust with third party medical apps, and <strong>WHY</strong> giving their data will help towards the affliction of currently incurable global disease.</p><p>Could it be that simple for other brands to follow this model? What would happen if brands began to be completely open about <strong>WHY</strong> they need to collect personal data, <strong>HOW</strong> they will be using it, and <strong>WHAT</strong> value this will bring to a consumers personalised service.</p><blockquote>A dedicated system to collect data +</blockquote><blockquote>Appropriate consent of what will happen with that data</blockquote><blockquote>= Where that data will be used to provide consumer value</blockquote><p><em>Terms and Conditions</em></p><p>Terms and conditions are a great example of a total failure in how they are designed for consumers to understand. The comprehensive detail that has to go into them boggles the mind, and its no surprise consumers can’t be bothered to read them.</p><p>Terms and Conditions are here to stay though, and as human rights laws become more and more transparent with what can and cannot be done with consumer data in the cloud, brands should consider measuring how satisfied consumers are with understanding the Ts&amp;Cs. <a href="http://www.fastcodesign.com/1665000/how-to-fix-the-nightmare-of-apples-terms-of-service">Greg Bernstein</a> explored this with a correction in redesign when concerning Apple’s iTunes T&amp;C oversight.</p><h4>The crisis of big data</h4><p>Is the crisis of big data amongst brands going to turn into a new commodity between governments and commercial brands? Much like what we have seen with the way Financial Services, Natural Resource and Food industries have operated to maintain political power, cloud-based data services are establishing a new gambling chip.</p><p>For the time being, brands are in a state of inertia as they continue to collect more and more consumer data to keep on trying to create that personalised service. A debate will certainly need to begin between governments commercial brands though, as change is almost certain to occur to make sure brands are being patient in the way they are identifying what conversations between consumers need to happen, and how brands will comply with the privacy and security governments will expect to protect the public.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4a8b7092559b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I hate the word ‘Accessibility’ — aren’t we designing experiences for humans?]]></title>
            <link>https://medium.com/@TimDaines/i-hate-the-word-accessibility-aren-t-we-designing-experiences-for-humans-dbee5f134ce7?source=rss-322ceac276c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/dbee5f134ce7</guid>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[human-behavior]]></category>
            <dc:creator><![CDATA[Tim Daines]]></dc:creator>
            <pubDate>Fri, 17 Jul 2015 14:43:49 GMT</pubDate>
            <atom:updated>2015-07-17T15:10:38.253Z</atom:updated>
            <content:encoded><![CDATA[<p>When I was 11 years old the population of Earth was 5.2 billion people. Today, its 7.1 billion people and in a world of digital revolution I find myself continually faced with the challenge of keeping up with human behaviours that surround all diverse backgrounds. To make things even more complicated, there are an abundance of buzzwords that continually adjust the tone of designing for human experiences, but until we really know what these mean its like popcorn in a pan…some pop…..some don’t!</p><p>Now there’s one buzzword you may hear flying around that more often than not sits behind the scenes of designing experiences for humans — ‘<strong>Accessibility</strong>’.</p><p>Personally, I hate the word! Yes thats right, you read this correctly. Not only is it labelling humans by putting their diversity to shame, but the sheer mention of it to a bunch of 20 year old designers and developers about to launch that mobile-app for retired holiday makers, either puts you in the camp of unpopular or someone speaking like a martian from outer space.</p><blockquote>But we are Lean and Agile, and will learn if accessibility is a problem after launching the product.</blockquote><p>Say what now!? A familiar response I’ve considered more as an excuse these days. I commend the refreshing sign of bravery, even when reasons for <em>‘I like it’,</em> are padded with all over the place questionable user research. Lean without considering human diversity — that’s a WINNER for all our users, right?!</p><p><strong>Stop blaming the designers and developers!</strong></p><p>Lets flip the coin on its head though. Is this apparent neglect by designers and developers the real route of the problem? Spare a thought for the designers and developers of this world who are creating human experiences to deliver awesomeness!</p><p>Typically, whilst the accessibility standards are there to guide us, one might be convinced that they somewhat resemble the same experience as reading the terms and conditions of an ecommerce website, or a credit card application, or world known search engines. I think you get my point. Although not written in a legalised way, the accessibility standards are presented in somewhat shall we say…….an uninviting manner for a designer, developer and even UXer for that matter. Yet when examined in more depth, the information that is the makeup of the standards constitutes…..wait for it…… designing for humans would you believe.</p><p>Wittingly or not, from my experiences working in various public and private sectors, designers and developers are automatically labelled as putting accessibility as an afterthought. Why do I think this? In a recent interview for a UX Architect role, I had come to conclude that this ‘label bashing’ perhaps stems from a different problem — how have user-centred design principles been communicated from the articulated vision of what the hell are we making here?</p><p>I was alarmed to say the least that the UX lead interviewing me hadn’t communicated this from the outset of starting with her team of designers and developers, and simply went on to say:</p><blockquote>This role requires another me. I need someone to keep asking why the designers and developers haven’t thought about the personas in the development of this product. I want someone who can keep going right to the point where you are putting the honey to one side and reaching to use the stick.</blockquote><p>Isn’t this an aggressive tendency of ‘Change Management’? Do the personas say this person is hard of hearing and visually impaired? No wonder this labelling of designers and developers exists.</p><p>Ultimately though, all of us designing experiences for humans need to be asking questions about this during our sprints or waterfall approaches (yes I know, some of us still have to use this method):</p><ul><li>What goals are we designing experiences for?</li><li>With so many users, have we forgotten whats important to all humans before launching a product or service?</li><li>Have we forgotten what the sound a user makes once we have designed an experience that addresses their pain points and frustrations?</li></ul><p><strong>Graphic Design to the rescue</strong></p><p>Perhaps there is an opportunity emerging to consider re-designing the standards. With so much evidence driving the importance of presenting digital to effectively communicate ideas to the humans we are designing experiences for, do the tenants of graphic design have much to offer to the accessibility standards?Is the charge of graphic design the recipe needed to combine the elements of the standards into a cohesive meaningful deliverable that resonates with developers, designers and UXers?</p><p>Presenting information can in itself be useless, unless there is a general understanding of how users will react to new information. Using graphic design to communicate the accessibility standards could change the way we all think about them and support decision making around what code must, should not do and what style tiles need to consider for visually impaired users. A similar approach is being designed around the myriad of government communications, such as Gov.uk, where cited improvements seem to suggest improved comprehension and a reduction in mistakes.</p><p>Could it be that simple? Aligning the accessibility standards with designers, developers and UXers through the use of graphic design, to increase readability and comprehension? One thing is for sure, if we are to be ignorant towards supporting and helping designers and developers to think about designing experiences for human diversity, ask yourself this before pre-launching a product — ‘what could possibly go wrong?’.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dbee5f134ce7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>