Update on position of AI on hype curve: it cannot dream

2 Replies

Decorative image of a flowerIt would appear that I was wrong in 2020 when I suggested that artificial intelligence was near the top of its hype curve [see ‘Where is AI on the hype curve?‘ on August 12th, 2020].  In the past few months the hype has reached new levels.  Initially, there were warnings about the imminent takeover of global society by artificial intelligence; however, recently the pendulum has swung back towards a more measured concern that the nature of many jobs will be changed by artificial intelligence with some jobs disappearing and others being created.  I believe that the bottom-line is that while terrific advances have been made with large language models, such as ChatGPT, artificial intelligence is artificial but it is not intelligent [see ‘Inducing chatbots to write nonsense‘ on February 15th, 2023].  It cannot dream.  It is not creative or inventive, largely because it is very powerful applied statistics which needs data based on what has happened or been produced already.  So, if you are involved in solving mysteries (ill-defined, vague and indeterminate problems) rather than puzzles [see ‘Puzzles and mysteries‘ on November 25th, 2020] then you are unlikely to be replaced by artificial intelligence in the foreseeable future [see ‘When will you be replaced by a computer?‘ on November 20th, 2019].  Not that you should trust my predictions of the future! [see ‘Predicting the future through holistic awareness‘ on January 6th, 2021]

Enduring, authentic, ancient and modern

Leave a reply

Decorative photograph across SevernTwo weeks ago, over a period of forty-eight hours, I visited four churches. An unusual event for me.  We travelled from Liverpool to Bristol one afternoon to attend a Thanksgiving Service the following morning for an extraordinary engineer and a lovely man, Eddie O’Brien.  The evening before the service, we stayed in a village pub in Oldbury-on-Severn and after dinner walked up the hill to the 13th century church dedicated to St Arilda.  It was locked so we strolled around the overgrown churchyard along a narrow mowed path and enjoyed the view across the Severn to Wales.  The following morning we drove into Bristol city centre to attend the Thanksgiving Service which was held in the Zetland Evangelical church.  The church was plain, unpretentious and packed.  The service was led by a retired pastor who preached with a gentle, thoughtful passion about Eddie’s life and its meaning.  I knew only one, possibly two, facets of his life: his professional life as an engineer and leading exponent of experimental mechanics; and his life as a student.  Eddie was twenty years my senior and thirty years ago I supervised his MPhil and PhD in experimental mechanics.  He was in his fifties and I was in my thirties – it was a challenge for both of us and we learnt from each other.  When he graduated he presented me with a copy of his PhD thesis that he had hand-bound in leather himself.  We left Bristol after the service and drove north across the Severn Bridge to Tintern Abbey where we stopped for lunch looking out over an empty cricket pitch across a green enclosed valley before exploring the ruins of the Cistercian abbey.  The abbey was founded in 1131 and in 1536 it was surrendered to Henry VIII during the dissolution of the monasteries.  The lead from the roof was removed and five hundred years of decay started creating the ruins you can wander around today.  Back in Liverpool, the following evening we went, with our neighbours, to a ‘Music at the Met’ concert at the Liverpool Metropolitan Cathedral called ‘Music for a King’ and featuring uplifting pieces, including ‘Zadok the Priest’ and ‘Crown Imperial’.  The bold grandeur of the concrete structure, richly coloured stained glass, thunderous organ and combined choirs of the anglican and catholic cathedrals contrasted starkly with the simple service of Thanksgiving for Eddie O’Brien we had attended the previous day when we sang hymns recalled from childhood, including ‘The Lord’s My Shepherd’.

Image: view across River Severn to Wales from St Arilda’s churchyard.

It was hot in June and its getting hotter

1 Reply

Decorative image of a summer flowerLast month was the first June on record when the daily average global 2-metre surface temperature exceeded 1.5 °C above pre-industrial levels [1] and last week, on July 6th, the daily global air temperature hit a record high of 17.23 °C [2]  In 2020 it was estimated that the world’s remaining carbon budget was about 500 gigatonnes CO2 if there was to be a 50% chance of limiting global warming to 1.5 °C. It is now estimated that the remaining budget is about 250 gigatonnes CO2, i.e., it has halved in three years, as a result of continued pollution and temperatures rising faster than expected [3].  At the current rate of emissions, this budget will be exhausted in about six years.  Hence, it seems very likely that global temperatures will rise by more than 1.5 °C and perhaps by as much as 4 °C this century.  The last time the Earth was that hot was about 15 million years ago during the Miocene when sea levels were 40 m higher [4].  It is time to get much more serious about reducing carbon emissions instead of just talking about it.  Current targets for reducing emissions are so far in the future that they are beyond the horizon – effectively out of sight and out of mind.  We need to be costing everything in terms of carbon emissions and making decisions that reduce emissions now.

[1] Climate graphic of the week: first days of June bring record heat, FT June 17, 2023.

[2] Global temperature hits record high, FT Weekend 8th July 2023 based on data from NOAA.

[3] Forster PM, Smith CJ, Walsh T, Lamb WF, Lamboll R, Hauser M, Ribes A, Rosen D, Gillett N, Palmer MD, Rogelj J. Indicators of Global Climate Change 2022: annual update of large-scale indicators of the state of the climate system and human influence. Earth System Science Data. 15(6):2295-327, 2023.

[4] Foster GL, Rohling EJ. Relationship between sea level and climate forcing by CO2 on geological timescales. Proceedings of the National Academy of Sciences. 110(4):1209-14, 2013.

Conflicted about cost-benefit analysis of international conferences

3 Replies

Decorative image of an aircraftLast week I wrote about my stimulating experience of attending a conference in Orlando, Florida and presenting our recent research to the experimental mechanics community for the first time in four years.  Whilst there, I was conscious of the ecological footprint of my trip – the venue was making extensive use of single use plastics on a scale that surprised me.  However, my trans-Atlantic flight had an order of magnitude larger impact.  It is difficult to find a reliable estimate of the carbon emissions for a return flight between the UK and Florida but 1,267 kg CO2 from the Guardian newspaper website lies between a lower bound estimate of 856 kg CO2 from iata.org and and an upper bound of 2,200 kg CO2 from myclimate.org.  This is equivalent to about one-sixth of my annual domestic carbon footprint of 9,000 kg CO2 using the calculator on the World Wildlife Fund website.  The UK average footprint is 9,300 kg CO2/capita and the global average is 6,300 kg CO2/capita.  The question is whether it is justifiable to generate additional emissions to attend a research conference?  The prime motivation of the research that I presented is to support the development of aircraft which are lighter with less embedded carbon and use less energy while also having a longer useful life.  Ultimately, supporting the aviation industry to achieve its target of zero-net emissions by 2050.  The carbon emissions of the global aviation industry in 2021 were 720 Mt CO2 [see IEA report]; hence, if my research contributes towards one hundredth of a percent reduction in these emissions then this would be 72,000 kg CO2/year.  It seems reasonable to cause a tenth of this annual saving each year (7,200 kg CO2/year) for the next ten years in order to deliver the required technology, i.e., committing one year’s savings to achieve an annual saving in perpetuity.  The problem is that I do not have a reliable estimate of the carbon footprint of my research activities.  I supervised an MSc student a couple of years ago who conducted a carbon audit of the School of Engineering and estimated the carbon emissions due to research alone to be 61,531 kg CO2 excluding heating, lighting and travel.  My group might be responsible for 10% of these emissions, i.e., about 6000 kg CO2; hence, adding about 1,200 kg CO2 to interact with other researchers at a conference seems reasonable and within a budget of 7,200 kg CO2. However, it is difficult to find reliable data to use in estimating carbon emissions for these activities and so perhaps the key conclusion is that we need more and better carbon audits to allow more informed decision-making.  In the meantime, perhaps attendence at an international conference once every four years is sufficient.

Image: Tayeb Mezahdia