Brandon Donnelly

Daily insights for city builders, delivered every morning at 6 AM ET. I’m Brandon Donnelly — a Toronto-based real estate developer and founder of Globizen. I’ve been writing here since 2013.

Subscribe

Tag: machine learning

  • Eating by algorithm

    Grocery shopping is one of those things that — despite a lot of people really trying — has remained a stubbornly in-person activity. However, the pandemic did give online grocery shopping a significant boost, and lot of that has stuck, even if it has been trending slightly downward from its peak. Here are a few slides from Dan Frommer’s Consumer Trends: 2024 Food & Wellness Special report:

    Part of the challenge may be that the majority of people say they actually like grocery shopping, and doing so in a physical store:

    So it is very possible that, for the foreseeable future, there will always be a large segment of buyers who prefer to shop in-store. But then again, if you asked me these same questions, I would also tell you that I like grocery shopping and that I prefer buying in-store. However, that doesn’t mean I wouldn’t be open to alternatives. I just haven’t explored and found a suitable online option.

    At the same time, and according to the same Consumer Trends Survey, about 10% of Americans say they currently dislike grocery shopping. Maybe this is the same 10% who are right now shopping online. Either way, this is already a large segment of people who would rather not go into a grocery store.

    Intuitively, as the online offerings get better, one would expect this number to grow. Here, for example, is an interesting overview of the service Hungryroot. One part “meal kit” delivery and one part online grocery shopping, the company uses machine learning and algorithms to determine what its customers might want to buy. Already, about 70% of what it sells is picked automatically.

    On the back end, McKean explains, among other actions, Hungryroot is “clustering” its new customer with other users who have answered its onboarding survey similarly and have already been with the service for multiple years. “And so we can say, ‘okay, people who filled out that signup flow like you… they loved these top recipes with high probability, so we think you’re going to love these recipes with high probability’.”

    What I like about this is that it requires fewer decisions; it has the potential to feel like you have a private chef (one that learns what you like and adjusts accordingly); and it promotes dietary variety. For the typical American, 75% of what they buy in a grocery store is the exact same as what they bought the last time. There’s very little variety, because it’s always easier not to have to think.

    Given this stat, it is maybe surprising that this 75% hasn’t become more automated for more people. Perhaps it’s the 25% that keeps most of us going into stores. I’m not sure, but I think I’m ready to try a service like Hungryroot.

  • The artificial intelligence bias

    Machine learning is one of the most important trends in tech right now. But like anything new, it naturally raises a number of important questions and concerns. Benedict Evan’s most recent blog post provides a good explanation of what he refers to as the artificial intelligence bias. Here are a couple of excerpts that I found interesting.

    What machine learning does:

    With machine learning, we don’t use hand-written rules to recognise X or Y. Instead, we take a thousand examples of X and a thousand examples of Y, and we get the computer to build a model based on statistical analysis of those examples. Then we can give that model a new data point and it says, with a given degree of accuracy, whether it fits example set X or example set Y. Machine learning uses data to generate a model, rather than a human being writing the model. This produces startlingly good results, particularly for recognition or pattern-finding problems, and this is the reason why the whole tech industry is being remade around machine learning.

    The rub:

    However, there’s a catch. In the real world, your thousand (or hundred thousand, or million) examples of X and Y also contain A, B, J, L, O, R, and P. Those may not be evenly distributed, and they may be prominent enough that the system pays more attention to L and R than it does to X.

    What AI isn’t:

    I often think that the term ‘artificial intelligence’ is deeply unhelpful in conversations like this. It creates the largely false impression that we have actually created, well, intelligence – that we are somehow on a path to HAL 9000 or Skynet – towards something that actually understands. We aren’t.

    The conclusion:

    Hence, it is completely false to say that ‘AI is maths, so it cannot be biased’. But it is equally false to say that ML is ‘inherently biased’. ML finds patterns in data – what patterns depends on the data, and the data is up to us, and what we do with it is up to us. Machine learning is much better at doing certain things than people, just as a dog is much better at finding drugs than people, but you wouldn’t convict someone on a dog’s evidence. And dogs are much more intelligent than any machine learning.

    Photo by Ales Nesetril on Unsplash

  • So the tech works. What’s next?

    Below is a keynote talk by Benedict Evans about what’s going on in tech today and what may happen in the next ten years. It covers: the growth of mobile; S-curves; Google / Apple / Facebook / Amazon (who knew Amazon had so many employees?); machine learning; autonomous vehicles/impact to cities; mixed reality; crypto-currencies; and so on.

    For those of you interested in crypto-currencies – and that appears to be everyone these days – it’s interesting to hear how Evans describes their current position at the beginning of the curve: “The tech works, but what’s the use case?” This is not to say the potential isn’t huge. It is. Automated trust. Distributed and programmable money. But the future is still unclear.

    If you can’t see the video below, click here.

    [youtube https://www.youtube.com/watch?v=cVYDkPidXrU&w=560&h=315]

  • Knowing how (automated medicine)

    I should read The New Yorker more often. I’m going to subscribe right now (done). The articles are fantastic. In the April 3, 2017 issue, Siddartha Mukherjee wrote a piece talking about automated medicine. It’s called A.I. Versus M.D.

    Here is an excerpt:

    His prognosis for the future of automated medicine is based on a simple principle: “Take any old classification problem where you have a lot of data, and it’s going to be solved by deep learning. There’s going to be thousands of applications of deep learning.” He wants to use learning algorithms to read X-rays, CT scans, and MRIs of every variety—and that’s just what he considers the near-term prospects. In the future, he said, “learning algorithms will make pathological diagnoses.” They might read Pap smears, listen to heart sounds, or predict relapses in psychiatric patients.

    This is a world where you could snap a photo of your unsightly rash, submit it, and then have a machine provide a diagnosis. This is a world where early detection and prevention would become common place. Already, today, there are instances where learning-based algorithms (distinct from rule-based) have proven to be more accurate than raw humans.

    I don’t know how you all feel about this, but I actually feel a sense of comfort thinking about this as the future of medicine. There’s piece of mind that comes with large amounts of data and the discipline of algorithms. At the same time, this is not necessarily about replacing humans. It is also about augmenting human ability.

    If it leads to better health outcomes, then why wouldn’t we?

  • Lists, disciplines, and AI

    I have a bit of an obsession with my calendar, lists, and goals. On the home screen of my phone I have Wunderlist, Evernote, Google Drive, and 2 calendar apps (more on that in a second).

    This obsession is probably one of the reasons I write this blog. I like writing, drawing, and documenting things. It helps me sort through my thoughts. I have everything from a list of cities I want to visit to a list of billion dollar business ideas (yes it’s really called that).

    But the other reason I like to keep lists is because – as I said in this post – I’m trying to be cognizant about overcommitment. And when I write things down, it serves as a reminder of what I’m doing and what I’m allowed to focus on.

    One of the ways I do that is through repeatable goals (or disciplines). These are non-negotiable things that I’ve committed myself to and that I just have to grind out – such as writing on this blog every day. I can tell you that some blog posts come out a lot easier than others. But I’m a firm believer that there are few substitutes for just showing up every day and putting in the work.

    On that note, let me tell you why I have 2 calendar apps on my phone. I’ve been testing out the latest version of Google Calendar because of a new feature they rolled out this month called Goals.

    What it does is automatically schedule repeatable goals. So for instance, you could tell it that you want to work out 3 times a week for 1.5 hours each time and that you prefer to work out in the evenings. It will then go and find 3 times for you to do it every week for all of eternity. 

    If a conflict arises, that’s no problem. As soon as you enter another appointment, the app will automatically reshuffle your calendar goals to make sure that you still get your 3 workouts in. It’ll even learn your preferences as you make changes to these appointments over time. 

    So far I’ve been finding this feature really useful. I used to do this for myself manually, but now I have a computer that does it for me. Even better.

    It’s also a glimpse into the future that Sundar Pichai, CEO of Google, was talking about on a recent earnings call:

    “We’ve been investing in machine learning and AI [artificial intelligence] for years, but I think we’re at an exceptionally interesting tipping point where these technologies are really taking off,” he said. “That is very, very applicable to businesses as well. So thoughtfully doing that externally we view as a big differentiator we have over others.”

    “In the long run, I think we will evolve in computing from a mobile-first to an AI-first world,” Pichai said in closing. “And I do think we’re at the forefront of developments.”

    For the past 5 years, it’s been all about mobile. But now much of the tech community – including the CEO of a pretty big company – is saying that artificial intelligence is next. What do you think this will mean for cities?