🗣 Attended a talk by Craig Larman entitled ‘AI and Org Design’ (kudos to AWA Global for making the recording of the talk available openly). Larman, co-founder of LeSS, has a long-standing academic and professional background in AI, and now describes himself as ‘junior GLAD developer’ (GLAD being an acronym for GenAI and LLM-assisted development).
In the talk, Larman demonstrated how even the current early, beta AI slash LLM tools are capable of replacing single specialism knowledge worker roles, roles characterised by routine expertise. As the tools advance, he predicts, they will make redundant many current roles, such as product managers, business analysts, UI/UX designers, frontend devs, database designers and their specialist managers.
Larman very politely debunked the wishful thinking indulged by many that AI will be added to our toolsets, making us more efficient. Even the current tools enable an order of magnitude increase in efficiency. Skill in doing a single thing will simply have no value. Instead, we will need skills in ‘multilearning’ (the concept goes back to the late 1970s and the work of Nonaka and Takeuchi), or the ability to learn and deploy new knowledge and skills quickly. In Nonaka and Takeuchi’s formulation, multilearning comprises learning across multiple levels (individual, team/group, and organisation), and across multiple functions. For Nonaka and Takeuchi, multilearning happens at the speed of change in the external environment, and in response to changing conditions. In Larman’s view, the rate of change is far faster, being driven by the rapid pace of development and adoption of AI tools. Consequently, speed of learning and adaptation at individual, team and firm level is all important. Larman quoted from François Chollet’s (the Google AI researcher) measure of intelligence paper:
‘The intelligence of a system is a measure of its skill-acquisition efficiency.’
In the Q&A Larman reflected on education, the focus in most public education systems on single expertise subjects, and the neglect of subjects which may encourage deep and wide understanding such as economics, psychology, philosophy, history and natural sciences, creative arts and humanities. Also this week, the UK Prime Minister announced the government’s intention to prevent universities offering what he calls ‘low value’ degrees, and this will likely mean that universities start cutting humanities and creative courses, in favour of single specialism vocational ones. University people sometimes get excited that they are ‘preparing students for jobs that don’t exist yet’. Now, if the UK government gets its way, we may find we are preparing students for single specialism jobs that won’t exist for much longer, when we should be enabling them to become multilearners for the future economy.
🔬 Since I last wrote, I have had a couple more conversations with ‘super learners’, individuals who are able to learn about new technologies quickly, while discerning the impacts and applications of these for their domains, and starting to adapt. I have watched as some of these people have become authorities, or ‘key people of influence’, in their domains (which so far include marketing, media and financial services), advising clients and speaking at industry events, in the space of 12 months or less.
As well as being voracious consumers of information (books, blogs, podcasts, journal articles), these people commonly have one or two learning partners with whom they regularly chat about developments and what they are learning, make sense, think aloud, reframe, challenge, test and refine their developing thinking. These learning partners are often colleagues present or past, business partners, friends, and they are hanging out and chatting informally and regularly, at least once a week, online and F2F.
🚵🏼♂️ Ride of the week: any mediocre effort of mine is unworthy of commemoration this week when we saw Jonas Vingegaard’s stunning performance on Stage 16 of the Tour de France, where he increased his lead over his nearest rival, Tadej Pogačar, from 10 seconds to nearly two minutes in the space of just 22 kilometres. Vingegaard said after the race: ‘I had really good legs today. It was one of my best days on the bike ever. I think at one point I started doubting my power metre because it was showing so high a number.’ Tom Dumoulin, former world time trial champion, said that Vingegaard’s ride was ‘the best time trial ever‘.
[This post features a banner picture showing a colour photo of a multi storey apartment block, eleven stories are visible, against a bright blue summer sky. Some of the apartments have bright orange sun shades over the windows. Location is Rotterdam in the Netherlands.]
Matthew Moran 