<?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 Dr Andrew Edmonds on Medium]]></title>
        <description><![CDATA[Stories by Dr Andrew Edmonds on Medium]]></description>
        <link>https://medium.com/@andy-edmonds?source=rss-319afd96f202------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*0PEWWDanLdIzmz11OYBwCg.png</url>
            <title>Stories by Dr Andrew Edmonds on Medium</title>
            <link>https://medium.com/@andy-edmonds?source=rss-319afd96f202------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 12:40:31 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@andy-edmonds/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[Musings on Democracy]]></title>
            <link>https://andy-edmonds.medium.com/musings-on-democracy-f0fd1e09d008?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/f0fd1e09d008</guid>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Sun, 31 Aug 2025 19:48:30 GMT</pubDate>
            <atom:updated>2025-08-31T19:48:30.830Z</atom:updated>
            <content:encoded><![CDATA[<p>I was in Switzerland last week, visiting Zurich. Switzerland is the world’s only direct democracy. It has a government and parties, but any decision of the government can be countermanded by a referendum, and new laws can be created by the same means.</p><p>If you’re Swiss, you just have to get enough signatures and your referendum question will be put to the entire citizenry in one of the regular referendum votes. The results are binding on the government.</p><p>You may have seen on the news that there were riots in Lausanne recently. Police were pursuing an immigrant youth on a scooter, who then crashed and died. Cars were burnt and policemen were attacked. Similar events have occurred in the UK and US for exactly the same reasons.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*akEzrWwKm9TjQ5nmLVrQAw.jpeg" /></figure><p>This was the first time I was in Switzerland for 20 years. Getting off a tram at the wrong stop had us walking through what turned out to be the wrong areas of Zurich. The Swiss themselves are even more neat-freaks than the Germans, and that’s saying something, but I was surprised how shoddy and dangerous chunks of Zurich were compared to my previous visits, and how few Swiss there were in these areas.</p><p>This led me to think about what differences there are between how a direct democracy and a representative democracy can handle the decay I saw and the rioting.</p><p>Whether or not the Swiss choose to fix the riots and the decay, they have the tools. Switzerland is not part of the spider’s web of international agreements that make lawyers rich and thwart the people. It is also incredibly hard to become a Swiss citizen: the rioters were almost certainly not Swiss. Finally, they have the referendum mechanism by which the people can force change.</p><p>In our representative democracies across Europe, we have a ruling class who do not want to fix the same problems, and a people who increasingly do want to fix them. The process of getting change in a representative democracy, where the ruling class do not share the opinions of the citizenry, is very slow. We may well be into civil war before the governments respond.</p><p>Wandering around Zurich and through train stations, one sees many young men in military uniform. Every male Swiss citizen has to go through military training. At the end they get to take their rifles home. It struck me this was very similar to Ancient Athens. When the Athenian army took to the field it was composed entirely of citizens. The middle classes could afford hoplite armour and so joined the main body of soldiers. The poor, known as <em>Thetes</em>, were skirmishers in the army and the rowers in the navy.</p><p>Athens had many foreigners, known as <em>Metics</em>, living there, as a great cultural and business centre, but none could vote. After the reforms of Pericles, getting automatic citizenship required both parents to be Athenian. <em>Metics</em> could be granted citizenship, but they had to be exceptionally useful to Athens.</p><p>The Swiss, by accident or design, have wisely followed the same pattern.</p><p>Away from Zurich, Switzerland is still a high trust society. We visited a nearby town called Thalwil, further along the shore of the lake, and saw a different pattern. I selected an ice-cream from a refrigerator on the street, and had to actively track down someone to pay for it. Nearby, expensive motorcycles leaned, with none of the complicated security now required elsewhere, and with helmets hanging off handlebars.</p><p>Now, if you are one of those “one world” people who think there should be no borders, the Swiss must appear very selfish. The Athenians lived in a time of constant strife, internal and external, and to limit citizenship, decision making and the bearing of arms to only those you could trust made sense.</p><p>There is another side to this, though. There has never been a stupider saying than “diversity is our strength”. In democratic societies, homogeneity is the superpower.</p><p>Democracy requires a “demos”, a citizen body. If this body has a common culture, then sharing is made much easier. If you are asked to pay higher taxes, but you are aware that the poor in your group, or your parents, or even you in your old age, would benefit, you are more likely to vote yes. If your country is composed of lots of different groups, you find squabbles for resources, “community leaders”, pork barrel politics and dissension.</p><p>A diverse society becomes a low trust society quickly, and low-trust and democracy don’t mix.</p><p>Interestingly, in both Germany and France, some parties’ manifestos have begun to include requirements for referenda on the Swiss model. Those who believe Germany’s AFD are neofascists would be surprised that this is on the first page of their manifesto. Giving decision-making to the people is as far away from fascism as you can get.</p><p>Switzerland, like many European nations, has experimented with large third-world immigration. Like most European countries it has discovered that this was a mistake, that over a certain threshold immigrants don’t integrate, they form ghettos.</p><p>It will be very interesting to see how a free citizenry navigates this problem in the future.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f0fd1e09d008" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The transgender debate illustrates what’s wrong with representative democracy]]></title>
            <link>https://andy-edmonds.medium.com/the-transgender-debate-illustrates-whats-wrong-with-representative-democracy-9701a9fd5981?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/9701a9fd5981</guid>
            <category><![CDATA[lobbying]]></category>
            <category><![CDATA[transgender]]></category>
            <category><![CDATA[direct-democracy]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Wed, 16 Apr 2025 16:00:54 GMT</pubDate>
            <atom:updated>2025-04-16T16:00:54.503Z</atom:updated>
            <content:encoded><![CDATA[<p>One of the major problems with representative democracy is the representatives themselves. They are always few in number, compared to the citizens, and while this theoretically makes decision making easier, it also makes ideological capture very simple.</p><p>In a direct democracy, you have to persuade half the population to make a change, in a representative democracy you just need to persuade a sufficient proportion of the representatives.</p><p>In the UK we have, so we’re told, a “Burkean” philosophy for our representatives. Edmund Burke proposed the idea that a representative should follow his conscience rather than that of his electors. It is considered “populist” to blindly just do what your constituents want. It is often forgotten that Burke, in his lifetime was only an MP for so called “rotten boroughs”, these being constituencies controlled by a powerful local landowner. Ignoring your electors therefore meant something entirely different in his day.</p><p>Trans people make up somewhere around one in a thousand in the population. The attention given to trans issues in schools seems to have increased this percentage by whatever mechanism, but there are still very few trans people. If we assume that gay people are all persuaded of the importance of trans issues, and this is far from clear, this still bumps up the percentage of those who actively want changes in the law to favour trans people to possibly 5%.</p><p>Somehow, though, trans ideology has been enormously successful in infiltrating universities, the media and politics. The general public, however, has not been persuaded.</p><p>I’m writing on the day that the supreme court in the UK has struck down the central tenet of trans ideology, that men can become women and vice versa.</p><p>My interest is not in the rights or wrongs of this issue, but how such a small group of people created such a strong movement, and what this says for democracy.</p><p>Clearly with only 650 people in the UK parliament, you only have to persuade 326 to get your way. This is certainly dramatically easier than the 33 million you might require in a direct democracy. This is not the only issue where representatives are at odds with the population. The climate change act, for instance, passed with only 3 dissensions. It is far from clear that this would have happened if a referendum had been held.</p><p>There is really only one country with a functioning direct democracy, which is Switzerland. There, the people have the opportunity to review any government law or decision, if sufficient signatures can be collected. The Swiss have recently voted for gay marriage, but against a climate change act. Switzerland has a mature democracy, and the people well understand their role. The bogeyman of “the tyranny of the majority” does not seem to be a problem there.</p><p>Interestingly, the Swiss government knows not to ever commit to anything the people would vote against. The EU recently tried to impose trade terms on Switzerland that the government knew the people would not accept. The Swiss government’s response was to walk away.</p><p>The idea of the right of the people to a review of government actions via referendum is growing across Europe. The AFD in Germany included it on the first page of their manifesto for the recent election. It was one of the demands of the “Gilet Jaunes” in France a few years ago. It would take only a small change to our system to make it possible, but would enable “the wisdom of crowds” to overcome the groupthink and embedded ideology of our representatives, and make it much harder for small but persuasive minorities, or lobbying billionaires, to take control.</p><p>One day soon we will realise all across the western world that a change in direction is needed, that the fixations of our elites have actively damaged our civilization.</p><p>The only way we can prevent a recurrence of similar malaises is to grant the people veto power over these elites. The time for direct democracy has come.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9701a9fd5981" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Things to beware in AI for businesses and developers]]></title>
            <link>https://andy-edmonds.medium.com/things-to-beware-in-ai-for-businesses-and-developers-13b51a5920e9?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/13b51a5920e9</guid>
            <category><![CDATA[ai-governance]]></category>
            <category><![CDATA[ai-compliance-issues]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Sun, 30 Mar 2025 14:01:17 GMT</pubDate>
            <atom:updated>2025-03-30T14:01:17.884Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*w96FqN-dL41GKpXE" /><figcaption>Photo by <a href="https://unsplash.com/@mrnuclear?utm_source=medium&amp;utm_medium=referral">ZHENYU LUO</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Artificial intelligence and Machine Learning have had enormous growth in the past few years. Their skills appear uncanny. Users are often left asking “what can’t they do?”, and the assumption is often made that there are no limitations, that AI can solve any problem.</p><p>I’m a veteran of AI. I started an AI company in the first wave in the 80s and have seen all the ups and downs in this field. Interestingly, the same questions I was asked with my first products are still relevant today: How do I know I can trust the results? Can I use it to predict the markets? And so on. The one new question is: Will this new AI product be subject to regulation?</p><p>I’ve observed that in recent years we’ve lost our natural common sense and caution about AI. We’ve become more credulous.</p><p>It’s important to understand the limits of AI and the potential pitfalls in assuming that any problem is solvable with a Large Language Model (LLM), or any form of AI.</p><p>Computer scientists have long known that some problems are, in practice, insoluble. For these problems you may be able to find a workable solution with a lot of computation, but you can’t find a perfect solution before the earth crumbles into dust. Other problems, like predicting the weather, climate, stock market, pandemics, earthquakes, politics, social media, etc., are provably mathematically intractable, and thus incomputable long term. AI programs are subject to the rules of mathematics too.</p><p>I have often seen companies persuaded to invest in solving problems that fall into these categories by the unscrupulous or ignorant. Success is possible, but very high risk.</p><p>The next problem with AI is that the big-tech LLMs, the dominant form of AI at the moment, are “black boxes”. This phrase is used for systems that create answers but can’t explain how they got to them. If, for regulated industries, you need to explain how a decision was made, you’ll have difficulty if you used a black box system. This is a specific requirement of the EU’s GDPR.</p><p>Another aspect of “black boxness” is that a question you ask of an AI system maybe one with loads of training examples when it was created, or very few. It can’t tell you which. When it’s few the AI makes stuff up, and this is the source of the “hallucinations” LLMs are prone to.</p><p>The big tech companies are very aware of the black box problems and are trying to find solutions. Currently this is work in progress.</p><p>Any system that learns relies on the quality of the data it learns from. It is also limited to giving results drawn from the data it was trained on. Noisy, sparse or contradictory data will result in bad outputs. In the world of LLMs questions about controversial subjects will illicit middle of the road answers that represent the zeitgeist at the time of training.</p><p>It’s important to understand that though LLMs appear to be creative, they are just assembling from a vast set of training data. There are other AI systems that are capable of true creativity, but LLMs are always just mixing and matching work previously done by humans.</p><p>Governments are regulating the use of AI. Regulations are, of course, political, and represent the concerns of the legislators who created them. The most complete set of regulations are in the EU’s AI act, but there are many regulations in other parts of the world, and though President Trump has recently foresworn AI regulation at a federal level, many US states have AI regulations at a local level.</p><p>If you are contemplating creating an AI solution, the fact that it might be regulated or banned in an important jurisdiction should give you pause. Even if you are creating AI enhanced software for internal company use, if you have a branch in an area with regulations you may have to limit the roll out. So, regulations in one important jurisdiction tend to creep into others.</p><p>Regulations are generally focused on applications. The EU, for instance, has a set of applications that are outright banned, and another set that must be regulated and licensed. Where licensing is required, there is often a testing regime.</p><p>Currently, as there is no case law, lawyers are advising AI developers to be super cautious. An example I recently heard of is a company spending seven figures to develop a system that monitors email interactions between customers and sales people, using AI to track sentiment. This was to be used in jurisdictions including the EU.</p><p>The EU’s AI act defines several “High risk” (effectively banned) applications of AI. The relevant one is:</p><p><em>“AI intended to be used for making decisions on promotion and termination of work-related contractual relationships, for task allocation and for monitoring and evaluating performance and behavior of persons in such relationships.”</em></p><p>The lawyers told the developers to abandon the project. I’ll leave you to decide if they were right or not!</p><p>The testing regime used where applications are licensed can also be problematic. Often one of the criteria is “equity” which is not a definable quality. In New York City, for instance, commonly used, AI based CV filtering software used by employers has to be tested every year for “equity”, by which they mean no racial or gender bias. Worthy as this might be, unless the CVs specifically contain the racial/gender characteristics of the applicant, achieving perfect equity is an impossible task.</p><p>Companies are racing to implement AI solutions everywhere, whether driven by a desire to automate, cut costs or just not miss out on an opportunity. In my experience, they almost never seek independent advice before they start. The results can be great, disappointing or disastrous. The latter two are much more common than you might think.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=13b51a5920e9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Top down versus bottom-up entrepreneurship]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-snippet">I&#x2019;m just in the middle of negotiating with an entrepreneur over some AI/Language related work. He wants to do something speculative, and&#x2026;</p><p class="medium-feed-link"><a href="https://andy-edmonds.medium.com/top-down-versus-bottom-up-entrepreneurship-accc9851e95e?source=rss-319afd96f202------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://andy-edmonds.medium.com/top-down-versus-bottom-up-entrepreneurship-accc9851e95e?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/accc9851e95e</guid>
            <category><![CDATA[invention-ideas]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Sat, 09 Nov 2024 15:14:58 GMT</pubDate>
            <atom:updated>2024-11-09T15:14:58.980Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Using AI to advise on AI compliance]]></title>
            <link>https://andy-edmonds.medium.com/using-ai-to-advise-on-ai-compliance-0fd0cf4fd84d?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/0fd0cf4fd84d</guid>
            <category><![CDATA[ai-compliance]]></category>
            <category><![CDATA[ai-regulation]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Thu, 02 May 2024 07:55:11 GMT</pubDate>
            <atom:updated>2024-05-02T07:55:11.258Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RIPWs0uta7RfhGKA7w8gWg.png" /></figure><p>Many jurisdictions are proposing to regulate AI, and some, such as the EU, have already done so, though the regulations don’t come into force ’til 2026.</p><p>Meanwhile a wave of startups are trying to use Generative AI and other forms of deep learning to do exciting things. Many of these will fail, of course, but some will fail because the thing they are trying to do becomes regulated or outright banned.</p><p>There are several fundamental problems with AI based applications that regulators will seek to control:</p><p>1. AI deciding things that impact an individual’s opportunities and choices.</p><p>2. AI deciding or predicting something without being able to explain how it did it.</p><p>3. AI making up plausible facts and passing them off as true.</p><p>4. AI recycling intellectual property without attribution or payment</p><p>Meanwhile, the makers of AI tools are lobbying hard to restrict regulation and in many cases working hard to obviate the problems, especially number 3.</p><p>The first consultation document from the EU focused heavily on (2), in proposing a “right to an explanation”. So, as proposed, any consumer who suspected they were denied something by AI, like a loan or cheaper prices on an online store, could ask for an explanation provided by a human. This has been dramatically watered down in the final AI act; I suspect as a result of lobbying. None of the most popular AI tools can explain how they arrived at any given decision.</p><p>Instead, the EU’s AI Act focusses on (1) and has banned some applications that might do this in a particularly egregious way, and required others to submit to a regulation process.</p><p>The dangers of AI are dependent on the application. If you want to use generative AI to write a story, by definition (3) is not a problem but (4) maybe, if, for instance, it writes a story where the hero is a young boy called Harry who lives under the stairs.</p><p>If a large finance company is suspected of granting more loans to one group of people than another, and they used AI to make the decisions, then (2) will be a major problem for that company if regulators investigate.</p><p>AI has been around for a surprisingly long time. The first attempts to create AI were made in the 1950s. In that time a very wide range of algorithms and techniques were developed.</p><p>In recent years, only one technique has received major attention, and that’s deep learning. Deep learning is a development and elaboration of neural networks, first proposed in the 50s, and with their first successes in the 80s. Generative AI is based on deep learning.</p><p>For many newcomers to AI, deep learning is all there is. It’s all they’ve been taught. But deep-learning derived models all suffer from (1) to (4) to varying degrees.</p><p>So, if you are an entrepreneur wanting to start an AI based enterprise, an angel investor or VC deciding whether to invest in an AI startup or a manager of an existing company planning to use AI tools with your customers or your workers, how do you navigate this complex landscape?</p><p>Well, you could ask a lawyer, who could help with interpreting the existing and proposed regulations. A lawyer who is also an expert on AI is a very rare beast though, and expensive if you can find one. There are many AI value-add companies, and the big tech companies all have their AI tools, but again the price for investigating a speculative idea would be very high, and the advice might well be partial. I’ve attended many AI presentations from big tech, and unsurprisingly the downsides are never mentioned.</p><p>I was asked recently to try to create an online tool to solve this conundrum.</p><p>It could be solved by using Generative AI, and indeed the first draft used this technique, feeding in the existing and proposed regulation to a RAG layer on top of an LLM. This version came out with nicely presented answers but hallucinated spectacularly, made regulations up, and couldn’t discriminate between proposed and existing regulation!</p><p>This is an application where you have to fall back on other forms of AI that aren’t black boxes, that are testable and auditable. i.e. a technique that doesn’t suffer from (2).</p><p>These techniques can’t do the fun stuff generative AI can, but they are perfect for this application.</p><p>My company, ThinkBase LLC, makes an eponymous tool that merges fuzzy rule-based systems and knowledge graphs. It doesn’t suffer from (2–4) and can be used more safely for (1).</p><p>We’ve created a dynamic knowledge graph that encapsulates both the existing state of regulation and the wider risks associated with using particular forms of AI. Once you’ve created such a graph you can reuse it either by presenting data in batch form, or where the system asks questions sequentially using graph theory to ask the minimum number of questions required to get an answer. We’ve used Microsoft’s Bot Framework to host this conversation, and the result can be used in Teams, on the web via SMS and many other routes.</p><p>You can try this out at <a href="https://aiethics.bot/">https://aiethics.bot</a> at a price way below the other alternatives.</p><p>For edge cases and further AI risk analysis we offer a human review, and if you want to do something that falls under existing or proposed regulation you should consult a lawyer to cover the fine print. We can help with that too.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0fd0cf4fd84d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The arbitrary nature of data analysis]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-snippet">When I first started my journey from electronic engineer to computer architecture designer to software designer, I carried the view I had&#x2026;</p><p class="medium-feed-link"><a href="https://andy-edmonds.medium.com/the-arbitrary-nature-of-data-analysis-b2e4c553628b?source=rss-319afd96f202------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://andy-edmonds.medium.com/the-arbitrary-nature-of-data-analysis-b2e4c553628b?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/b2e4c553628b</guid>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Wed, 14 Jun 2023 13:35:31 GMT</pubDate>
            <atom:updated>2023-06-14T13:35:31.610Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Why do people let their enemies define them?]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-snippet">The press holds some stories today about Elon Musk criticising George Soros in various tweets. What&#x2019;s interesting is that several bodies&#x2026;</p><p class="medium-feed-link"><a href="https://andy-edmonds.medium.com/why-do-people-let-their-enemies-define-them-3cc8df2c0530?source=rss-319afd96f202------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://andy-edmonds.medium.com/why-do-people-let-their-enemies-define-them-3cc8df2c0530?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/3cc8df2c0530</guid>
            <category><![CDATA[elon-musk]]></category>
            <category><![CDATA[george-soros]]></category>
            <category><![CDATA[race]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Fri, 19 May 2023 08:23:40 GMT</pubDate>
            <atom:updated>2023-05-19T08:23:40.661Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Meditations on Chaos and Order]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://andy-edmonds.medium.com/meditations-on-chaos-and-order-6bdab04cdd74?source=rss-319afd96f202------2"><img src="https://cdn-images-1.medium.com/max/1078/1*lE06sadYRy3IApLpPEtAzw.jpeg" width="1078"></a></p><p class="medium-feed-snippet">This is a picture out of my window on a rainy May day.</p><p class="medium-feed-link"><a href="https://andy-edmonds.medium.com/meditations-on-chaos-and-order-6bdab04cdd74?source=rss-319afd96f202------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://andy-edmonds.medium.com/meditations-on-chaos-and-order-6bdab04cdd74?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/6bdab04cdd74</guid>
            <category><![CDATA[predictions]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[chaos-theory]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Fri, 05 May 2023 14:31:16 GMT</pubDate>
            <atom:updated>2023-05-05T14:33:18.369Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[The blindingly obvious ignored, again]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://andy-edmonds.medium.com/the-blindingly-obvious-ignored-again-dfe4eae8f12a?source=rss-319afd96f202------2"><img src="https://cdn-images-1.medium.com/max/1600/1*MU1Ed-8aP2OevhsP-anrdw.jpeg" width="1600"></a></p><p class="medium-feed-snippet">This piece will tackle two subjects, either of which would get me cancelled forever, were I a user of social media.</p><p class="medium-feed-link"><a href="https://andy-edmonds.medium.com/the-blindingly-obvious-ignored-again-dfe4eae8f12a?source=rss-319afd96f202------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://andy-edmonds.medium.com/the-blindingly-obvious-ignored-again-dfe4eae8f12a?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/dfe4eae8f12a</guid>
            <category><![CDATA[race]]></category>
            <category><![CDATA[solipsism]]></category>
            <category><![CDATA[gender]]></category>
            <category><![CDATA[truth]]></category>
            <category><![CDATA[culture]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Fri, 11 Mar 2022 12:56:48 GMT</pubDate>
            <atom:updated>2022-03-11T12:56:48.176Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[We need direct democracy, and we need it now]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://andy-edmonds.medium.com/we-need-direct-democracy-and-we-need-it-now-202253ce97fd?source=rss-319afd96f202------2"><img src="https://cdn-images-1.medium.com/max/600/1*Nnoc8lSBV-OZcad7cFofJg.jpeg" width="600"></a></p><p class="medium-feed-snippet">In my last article I defended Jordan Peterson, in this one I&#x2019;m going to disagree with him. In a recent podcast he said, and I&#x2019;m&#x2026;</p><p class="medium-feed-link"><a href="https://andy-edmonds.medium.com/we-need-direct-democracy-and-we-need-it-now-202253ce97fd?source=rss-319afd96f202------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://andy-edmonds.medium.com/we-need-direct-democracy-and-we-need-it-now-202253ce97fd?source=rss-319afd96f202------2</link>
            <guid isPermaLink="false">https://medium.com/p/202253ce97fd</guid>
            <category><![CDATA[history-of-politics]]></category>
            <category><![CDATA[jordan-peterson]]></category>
            <category><![CDATA[direct-democracy]]></category>
            <dc:creator><![CDATA[Dr Andrew Edmonds]]></dc:creator>
            <pubDate>Sun, 20 Feb 2022 15:59:46 GMT</pubDate>
            <atom:updated>2022-02-20T18:01:54.600Z</atom:updated>
        </item>
    </channel>
</rss>