2023-12-24

I may have missed my weekly journalling slot last week, I had gotten in the habit of simple using the time when the big guy was at basketball practice to sit down and pen some thoughts on my phone. However, with the end of basketball, I suddenly had a Sunday without the regular routine. I had all the intentions of sitting down and still finding some time to write, but then life happened. Although reflecting back upon it now, I’m not quite sure what was more important. I seem to recall that my parents were off on their wedding anniversary weekend, so I didn’t even have my regular Sunday Skype with them.

So what did I do with last Sunday?

Who knows! Obviously it wasn’t that memorable in the grand scheme of things.


Update: I’ve just recalled what it was that I did with my last weekend. I had a very sick toddler with a temperature bordering 40 for the whole weekend. And when the little one gets sick, he becomes like a shadow. Attached to me at all times. Saturday night, as I climbed into bed, I could feel a phantom toddler sat still pressing into my thighs from where he’d been sat all day.

On Monday morning, in the middle of a thunderstorm, I took him to the paediatrician to confirm that he was mostly fine and/or seek antibiotics. On Friday night he came home with some gastro, but that seemed to evolve into a high-fever ear infection. The good doctor, seemed more concerned about the recent cough and rash. She went on to explain that the city is in the gripes of a nasty gastro bug, mumps outbreak (because of stupid anti-vac mums) and a scarlet fever outbreak.

Thanks to us now living in a post-Covid world, apparently there are now rapid tests for everything. So a quick scrape of the throat with the nurse and a 15 minute wait, and we found that he was negative for scarlet fever, and we were on our way.


Xmas shopping

What I do recall of last weekend was taking the big guy downtown to do some Xmas shopping. Not because we need more detritus in our lives or in the apartment, but because we feel that it is important to continue nurturing the habit of getting him to think of others.

As we walked down the hill towards the shops, I struck up conversation with the big guy - trying to prepare him for what was to come. Trying to make him of things that the other people in our lives might like as a present.

Automatically, he suggested that his mother and grandmother should get dresses. A little gender stereotyped perhaps, and, well, you know, it was -4 degrees out so I wasn’t sure how much wear a dress would get in the immediate future.

The point on contention that I raised with him however was that I didn’t know what size his grandmother. He replied, “Well how old is safta?”.

“I don’t know, she’s about 60, maybe 62.”

Without skipping a beat, “Then just get her a size 64 and it’ll fit.”

Hilarious! Because obviously in his world, he is aged 7, and the clothes he wears are all size 7 or 8. Extrapolating out, a 60 year old should obviously wear a size or two bigger than her age.

We proceeded to the local department store amongst the Xmas crowds and started our way through the accessories section. It’s odd what kids focus on and what they get obsessed by. For instance, when we walked into the store, he saw an umbrella with a transparent canopy and right away was convinced that this was the one thing that was missing in his mother’s life. Despite, three attempts to move him onto something else in the store, each time we looped back around until we stood in front of the umbrella stand.

As is traditional for his gender, we mostly blitzed through the store picking up items quickly for every member of the family as if we were on some sort of competitive scavenger hunt. Some sunglasses here, some objective awful gold hoop earrings there. Then we could get down to the proper business of finding him some coloured pens and toys in the bookstore.

The work run to Xmas

On the work front, I was very much looking forward to a quiet run into the break. It feels like this stint at work has been pretty long, and although the fall season has been pretty forgiving in terms of weather and adjustment to limited daylight hours - I am still very much in need of the break.

Way back in September, the current client gave us an ambitious target of creating ten use-cases to prove the business value of the product and to have something tangible that could go into a release by the end-of-the-year.

My old school project management brain kicked in, and quickly did the calculation. Well I don’t want to work the break, minus 2 weeks. Let’s have a bit of buffer, for overruns, testing and bugfixes, minus another 2 weeks. That means the deadline was really December 8th. How many weeks between now and then? 13. Okay so that gives us a couple of weeks to get the process up and running with the first use-case and then it’s pretty much exactly one per week for the rest of the year.

Not an impossible challenge, but one that would definitely stretch the team.

As luck would have, very little went wrong with the original plan, and we had 7 use-cases complete by the December 8th deadline, with the final 3 nearly there, but just requiring a little finalization.

When all was said and done, the team essentially had one week spare, making for one of the lowest drama releases I’ve ever witnessed. Borderline boring and uneventful.

The complication came, however, when it came for the client product owner to actually demo the product to the program executive and to a broader array of execs from across the organisation. Our exec was looking for wider buy-in for a bunch of unimportant political reasons. But when the demo was conducted from an environment that was 6 months old, had none of the new futures, but also was routinely timing out as a bunch of the old features had been reworked, it resulted in a catastrophic demo, the likes of which I’ve never seen.

Of course, as an external consultant, at an internal meeting, I was there out of courtesy with a strong remit of being seen and not heard. Which meant that I got to witness the train wreck in slow motion without being able to intervene in any way.

Soon after, the program exec was on the blower demanding a post-mortem and retrospective on the previous year of efforts.

It’s not exactly the position you want to be in as an AI consultant, where the client is demanding to know exactly where is the AI in this alleged AI solution that they’ve paid millions for.

Nor do you ever want to be in the situation, where anyone is asking you, “What exactly did you do last year?”

The week therefore devolved into emergency meetings, digging through slides, statements-of-work, quarterly planning minutes, and then building of a narrative.

Crafting a narrative that we’ve actually done a lot, without sounding defensive, when the client suspects that you’ve done nothing and is starting to suspect that all of your reporting is increasingly suspect.

My opinion is that you strip everything back to basics. Without the marketing fluff. Without the polished slide decks. Go back to an old fashioned demo of the product. Because what we’ve done is impressive, and it does provide value. But if all you ever do is present abstract slides, that feel like marketing hype, it’s hard to keep track of the actual product.

Especially in this case.

In this case, there are actually 3 AI modules, each comprising several sub-components, each with their own progress, results and tuning. As a guy running the show, day-to-day with nothing else to focus on, it’s a big ask.

To expect a bunch of executives, that are engineers first, and certainly not statisticians or software developers to keep all of these intricacies in their head on a week-to-week basis - in retrospect is stupid.

Maybe that’s my takeaway from all of this. When we go back to basics, and talk about the AI in terms of “cats and dogs” (i.e. an image classifier that has been trained on a library of dog pictures), then everybody can keep up.

But I should know this by now. Execs are problem solvers. There is an expectation that when things are brought to them, they are there to generate ideas, make decisions. I’ve joked in the past that when taking things to senior management, their first reaction is to always jump on things like it’s a live grenade. There’s an assumption, that if things are being raised to their level, then it requires their immediate attention.

If it doesn’t. If it truly is for reporting purposes only, then that context needs to be incredibly explicit up front.

Now that I’m reflecting upon it, this is exactly the same problem we had this time last year with this client. The perception, true or not, was that we’d been burying the bad news under superficial reporting slide decks that masked the bad news that was really coming when the deadline was missed.

In the forthcoming new year, maybe I should involve myself in making sure that some reality is reflected to the client. Force some demoes. Relate everything back to “cats and dogs”. Not just get caught in the trap of glossy slides.

This time last year, was the worst release process I’ve ever worked through. Weekend standups. 7 AM alignment calls, 7 PM alignment calls, and then the worst client meeting of my life.

This week, it wasn’t quite the same. It wasn’t the worst client meeting of my life. But it certainly was the most stress I’ve felt in a long time. You know it’s bad when executives are calling their sectaries to cancel their next couple of meetings. You know it’s bad when an executive call breaks for lunch and then keeps going for another 90 minutes in the afternoon.

Let’s not do that again.

Newer post

2024-01-11

What distinguishes you from other developers?

I've built data pipelines across 3 continents at petabyte scales, for over 15 years. But the data doesn't matter if we don't solve the human problems first - an AI solution that nobody uses is worthless.

Are the robots going to kill us all?

Not any time soon. At least not in the way that you've got imagined thanks to the Terminator movies. Sure somebody with a DARPA grant is always going to strap a knife/gun/flamethrower on the side of a robot - but just like in Dr.Who - right now, that robot will struggle to even get out of the room, let alone up some stairs.

But AI is going to steal my job, right?

A year ago, the whole world was convinced that AI was going to steal their job. Now, the reality is that most people are thinking 'I wish this POC at work would go a bit faster to scan these PDFs'.

When am I going to get my self-driving car?

Humans are complicated. If we invented driving today - there's NO WAY IN HELL we'd let humans do it. They get distracted. They text their friends. They drink. They make mistakes. But the reality is, all of our streets, cities (and even legal systems) have been built around these limitations. It would be surprisingly easy to build self-driving cars if there were no humans on the road. But today no one wants to take liability. If a self-driving company kills someone, who's responsible? The manufacturer? The insurance company? The software developer?