Weeknote 2020-15

Quite a productive week this week. I think our team is starting to work out what our place is in this current online-only environment. We are used to running in-person workshops and so it has been a challenge working out how to move them online, and yet still offer something different from the usual run-of-the-mill online offering already out there.


I attended two meetups/conferences this week:

COVID-19 and AI: A Virtual Conference, hosted by the Stanford Institute for Human-Centered Artificial Intelligence (HAI). This was a full day of talks, in a traditional 'conference' format. Divided into four sessions:

  1. Opening Remarks, Landscape and Framing
  2. Social Impacts & Bio-Security
  3. Tracking the Epidemic
  4. Treatments & Vaccines

I think some of the speakers were feeling a bit out of place doing their talks remotely, and I felt a lot of them were just reading a pre-prepared text they had. Whilst that sometimes works in-person, it very much doesn't add anything online.

There was disappointingly actually very little AI in the content. Much of it was just looking at policy and stats. That said it was very interesting talk by Kate Starbird talking about the spread of misinformation during a pandemic. And a great talk from Jason Wang on Taiwan's use of data analytics to track and control Covid-19.

But overall the format of 30 minute speaking slots followed by a panel discussion with all speakers seemed to work pretty well.

I also “attended” a BrisTech meetup in which Warren Boult, and engineer at Candide spoke about Kubernetes-native ML pipelines with Argo and Seldon. This is an area I think I'm going to need to be looking at more in the near future, as I think one of the next challenges is deploying ML models and training.

Creating Content

I great post shared by a colleague on “How Content Creates Content”. The post talks about how you can re-use and re-package various bits of content in multiple formats and for multiple attention spans.

This is a format I'm hoping to try and use myself going forward and be a bit more planned in my content and how it all relates to each other. So for example, I'm going to be doing Twitch streams on certain topics, and then write up a blog post about the topic covered. Those topics will then be used to generate in-person (online currently) workshops in which we go into more depth. And then those workshops used to generate tutorials, for example.


So yesterday I did my first live Twitch coding stream! Whilst I've presented live shows before, this is the first time I've coded live. And it all went pretty well.

There was a panic at the start as due to a communication mix up I didn't get the key to the streaming account until literally the minute I was due to start. And apparently my webcam was lagging a bit, but the screenshare and audio seemed to be in sync.

The fans on my Macbook were going like crazy, and I had to put a noise cancellation filter on the audio as you heard the fans in the mic otherwise. I was using an AKG Perception 120 USB mic that I've had for a while, and a Logitech C925 webcam. The quality seemed to come out OK, although I think the light from my windows was causing a bit of a white-out on my face. I'm due to be getting a green screen at some point, but for now had my trusty triffid in the background. We worked out that that plant is only a few years younger than the youngest member of my team!

I was demonstrating Watson Studio and pulling in some live data from Bristol City Council air quality monitors to see if we could see a change in the level of pollution compared to before the lockdown. To find out the answer (you can probably guess!) stay tuned for a dedicated blog post in which I walk though it.


I've always been aware of pronouns and what people want to be referred as. My eldest two kids are both transgender, and so it is something I've become very much accustomed to with both them and their circle of friends. But having said that, I've not really bothered with the listing of pronouns I want people to use for me in my bios on sites. Why? I'm a white, cishet, male... I'm the 'default', right?

Well yesterday I saw something that changed my mind. It was a combination of two things. Firstly a colleague of mine, who I only just 'met' (online) I would have assumed uses female pronouns has listed their pronouns as (they/she). So it prompted me to use the pronouns they want me to use without asking... but secondly and more importantly, we have a new CEO at IBM, and with that we have a new President too (moved over from Redhat). And in the first communication the President sent around, at the bottom, without any fanfare was this:

He has listed his pronouns. This was picked up on the IBM LGBT & Allies slack channel and the response was overwhelmingly positive. One particular comment stood out:

I’ve been following Red Hat for about 10 years and I’ve heard a lot of positive things about Jim. This is the first time I’ve had any communication directly from him though, so it counts for a lot

The small things. They matter. So why does it matter that people like him, or I list our pronouns if they are 'as expected' from our names and appearances? Because it normalises the practice. Five or ten years ago, if my colleauge above had listed their pronouns it would have seemed 'odd'. It would have been an abnormality. If the President of the company does it, then it is just one of those small subtle signs to others. That they are welcome, regardless.

So I am starting to update my bios to include my pronouns. 🏳️‍🌈