Tera Allas CBE

AI not all 'doom and gloom' for graduates, say businesses Divine Jacob feels AI creates extra barriers for job hunters but analysts say the use of AI is an "evolving picture".

Really pleased to have contributed to this BBC article (and programming). It's not all doom and gloom. Longer-term, there are uncertainties and risks, and the job market is in a transition, but right now there are high quality roles available for AI-literate candidates. www.bbc.com/news/article...

The Spring Statement 2026 was indeed a fiscal event with zero new policy measures and, in total, the smallest number of policies costed since 2014 [when I started tracking this]. Let's hope it is a start of a trend, towards fewer but more material reforms. #dataisbeautiful

Superintelligence is already here, today It's going to revolutionize science. It also might take control of this planet.

"We are witnessing a fundamental shift in the scientific workflow." The social contract with academia and researchers is about to break. This is an opportunity: how do we caerfully reshape it to maximise better outcomes? substack.com/@noahpinion/...

#ai #productivity #dataisbeautiful | Tera Allas CBE The flipside of #AI not showing up in #productivity is that it's also not showing up in the labour market. Demand for AI-exposed jobs has increased. The (modest) decline for typically young graduate j...

A slightly longer write-up about the chart above is here: www.linkedin.com/posts/teraal...

The flipside of #AI not showing up in #productivity is that it's also not showing up in the labour market. Demand for AI-exposed jobs has increased. The (modest) decline for typically young graduate jobs is public sector driven. Read productivity article here: mck.co/4b8w4wh #dataisbeautiful

And: I sympathise! Why is there an explosion in ‘must read’ stuff exactly at the point at which a) one is already too busy or b) one desperately needs to chill out?

Well, the latest Matt L (“AI can manage your mutual fund”) is an absolute must read :-) Laughed out loud because of its cleverness several times (even tho not unusual with him).

UK's #productivity growth might finally be picking up - but not particularly in sectors impacted by #AI. So what is going on and why is AI not yet boosting productivity (at the sector/economy level), despite its enormous potential? Read the full story here:https://mck.co/4b8w4wh #dataisbeautiful

Yes, I think we will. Maybe not in 2026, but probably in 2030. The AI we have today is the worst AI we'll ever work with. And in many, many applications customers will be willing to buy a worse product at, say, 10% of the historical cost.

If #AI is boosting UK #productivity, it's not yet showing up in the statistics. This is not a surprise, given the 'ladder of leakage': AI benefits leak out at every stage of aggregation, from tasks to teams to firms to sectors to the economy. Full article here: mck.co/4b8w4wh #dataisbeautiful

Thanks! AI is now massively better than it was a year ago, though, so the _technical_ potential is enormous. The actual gains are all about people (psychology, sociology, management) so lots of barriers to realising them.

Quite right - I enjoy baking and knitting, and still do both. However, I would not be able to compete in the market place with those skills. Pure coding may soon not be remunerated by the market place, either. Sad as it is, it may become a hobby and an artisanal craft that doesn't support a living.

Link to now published article: www.mckinsey.com/uk/our-insig...

I too need to think more carefully! But doesn't human labour only find uses if someone has money to buy the outputs? And if the marginal cost of nearly everything is the cost of compute, then labour can hardly compete. Won't nearly all income accrue to those who own resources (land, mines, etc.)?

Sure, their “ghost GDP” is nonsense, but the scenario where no-one but a few tech bros have jobs/income/wealth is still one to contemplate… Would democracy collapse or correct? Would the tech bros realise it’s only cool to be the richest/most powerful, if other lesser people are still around?

#ai #dataisbeautiful | Tera Allas CBE #AI adoption drives productivity growth, right? Just see the chart below (R2 = .37, p < 0.01)! Oh but wait - have a closer look... AI-exposed sectors' productivity grew faster BEFORE the current AI wa...

In the meanwhile, a slightly longer explanation here: www.linkedin.com/posts/teraal...

#AI adoption drives productivity growth, right? Just see the chart below (R2 = .37, p < 0.01)! Wait - have a closer look. AI-exposed sectors' productivity grew faster until Q1 2020, pre-ChatGPT. Check out the real story in our upcoming article (will post link here when live). #dataisbeautiful

The image shows three speakers in an event discussing "What can we learn about NHS productivity from developments across the wider economy?". The speakers are identified as Tera Allas CBE, Sir Charlie Mayfield, and Professor Bart van Ark.

WEBINAR: What can we learn about NHS productivity from developments across the wider economy? Chaired by @teraallas.bsky.social from our #NHSProductivityCommission, we will hear from an expert panel on what we can learn from sectors and industries beyond health care.

Register ⬇️
bit.ly/3ZJCRpV

Today's ONS release on healthy life expectancy (HLE, bit.ly/4cz4rxB) makes for grim reading. Since 2011-13, public expenditure per capita on health has increased by 25% in real terms, but HLE has dropped, and more steeply for women than men. #dataisbeautiful

Hot on the heels of our first evidence event on learning from other health systems, the #NHSProductivityCommission is holding its second webinar on learning from the wider economy. Join us for this session chaired by @teraallas.bsky.social w/ panellists @bartvanark.bsky.social & Sir Charlie Mayfield

I feel sorry for Claude Code having to deal with ONS nightmare spreadsheets :-( I mean, rather him than me, and for a while I thought the nightmare spreadsheets problem was 'over' because Claude was so obliging. But now I can see it's using up a lot of my valuable tokens!!! #dataisbeautiful

Unfortunately, the expected #productivity #growth revival is rather hard to detect so far... Even the RTI experimental statistics suggest growth stalled in Q4 2025. On both measures, output per hour is lower than Q2 2021... Not great for living standards. #dataisbeautiful

UK #unemployment rose to 5.2%, with an increasing gap between male (5.7%) and female (4.7%) unemployment. Many factors are driving this, but one is weakness in sectors with traditionally more male workers, such as construction and transport and storage. #dataisbeautiful

Fab—have been wanting someone to dig into this. But I have to slightly object to that pie chart… Not sure it matches the message.

Census 2031 might not measure wellbeing or life satisfaction. If you think it should, we only have a few days to act. The consultation closes next week (20 February 2026). Here's how to respond: bit.ly/4aBsERt

Is the drop in UK business investment in Q4 2025 a cause for concern? Well, we need the more detailed data to know. For now, I'm reassured that (whole economy) investment in intellectual property (incl. R&D) and ICT equipment and other machinery continued to increase. #dataisbeautiful

So "services" made very little contribution to 3-month #GDP growth in Q4 2025. But does it really make sense to lump together accounting and cleaning, retail and computer programming? The "services" label is pretty meaningless. #dataisbeautiful

One of the more striking findings of this excellent MGI report (bit.ly/4cth6lA) is not that demand for AI skills is rising but that demand for people leadership, management, and operational skills is rising. Good AI adoption requires good management! #dataisbeautiful

The UK labour market has slackened considerably in the last year - but it would be really hard to conclude this has anything to do with #AI, at least at the aggregate level. Vacancies and payrolls decreased most in not-very-AI-exposed-sectors. #dataisbeautiful

A wonderfully written synthesis of the history of causal inference in economics. Particularly enjoyed the analysis of where #AI may help, or not. "The history of econometrics suggests that methodological advances often outpace the understanding needed to apply them responsibly." bit.ly/4cpiLsm