It’s a Big, Big, Big, Big Data World
So I visited the Big Data LDN expo in London this week, and I actually enjoyed it. There, I really wrote that. Generally, I am not sure I am a good expo delegate: I do not go there to close massive deals (many of the exhibitors, or better, their products I already know) and receiving a plain sales pitch with depth quickly bores me.
And, of course, it was my first real expo since Covid (I will write about my experience with virtual conferences and expos another time). And yes, it felt odd when I stayed in the (really long) queue.
As a rule, I prefer to talk to delegates from smaller companies, as they are closer to the source, much more so than a hectically trained sales representative.
But let’s start with some impressions: Big data is around for a while now, the topics around it have not changed much over the last decade. The shear size of the data to deal with is still increasing as fast as hardware (and software) capacity. Making sense of all this data is still a challenge. Hardware however has been replaced by “the cloud”, so we stopped mostly talking about it.
What fascinates me that there was — despite the massive hype of the last years — much less AI or ML related products than I expected. That feels odd, considering that big data is handled by machines and machines alone, and nowhere else currently does practical application of machine learning make more sense.
While many exhibitors had some kind of reference to machine learning in their product description, I found very very few specifics: One exhibitor raised my interest with a NLP application (NLP being my personal interest for decades), but when asked what models and libraries this is based on, I got nothing but that “the models are their own”. Ah well.
I enjoyed the one event by Itoro Liney about AI ethics and machine learning bias, not because I learned much new (it was targeting a general audience, obviously, not specialists), but I really appreciate every attempt to raise awareness of the issue. And I do strongly believe — like with data privacy — AI Ethics needs to be discussed with everyone involved, from data scientist and developers, product managers to management.
The talk was in fact the only reference to machine learning bias I came across, which I find rather worrying for an expo of this size. While data privacy has widely been acknowledged as a topic in the industry since the introduction of the GDPR, issues with machine learning bias and ethics in data science has certainly not. Let’s hope that the EU’s AI Act — as limited as it may be — may do the same for AI ethics what the GDPR did for data privacy.
Ops, Ops Everywhere
And then there are Ops. DataOps, DevOps, ML Ops, AI Ops, Ops Ops. I have been told that whenever I hear this term a frown shows on my face. Approaching a stands featuring “Ops” leads me invariably to approach with a straight “What does it actually do?”, just because I can’t help myself. At least twice I got a vague “it can do everything you want with data” (which I doubt it can) as an answer. Ah, well.
In general, the trend to put things on top of other things continues here: An Ops layer on top of another ops layer on top of some third part software stack on top of a very specific cloud service. I am sure another all encompassing form of ops (“OpOps”?) is just around the corner.
Don’t get me wrong, Ops play a very important role, but nowhere else I feel more overwhelmed by buzzwords and use cases where I feel I model my operations after the service, rather than the other way round.
On a positive sidenote, I can happily report that the days where everyone had to clobber some mention of blockchain into their pitch appear to lay behind us. I for one won’t miss it. Not that I doubt blockchain technology can be put to good use, but let’s be honest, things got a little out of hand.
The Fun part
As mentioned, I prefer to approach smaller companies, and I think more space should be dedicated to those. I understand expos make the big bucks with enterprise, but still. I had some really good talks at those stands, listened to really good use cases and unique selling points, and met some really engaging people. Just wish there would be more time and space to network.
