Letters to the editor: volume 17, issue 5

Let’s learn more from the planet’s best computer

I would like to fully agree (well almost) with David Zilli’s suggestion that artificial intelligence needs to work smarter, not harder (Comment, April 2022).

The most successful project on the planet must be the six-million-year-old development of a carbon-based computer, the human brain. In all projects, there is normally a lessons-learnt phase. Artificial neural networks stem from a concept formed over 60 years ago. Only recently, in around the past 15 years, has computing technology been able to make good use of the AI concept. Feeding a neural network lots more training data does not necessarily make it smarter. It will start to identify coincidences and draw parallels that are incorrect, or, worse, solutions will be misleading.

Isn’t this rather like giving a child lots of food to consume, expecting them to grow up healthy? I believe the neural network concept has not gone far enough. AI does not need to be fed more training data to work smarter, AI needs a mentor. AI needs a teacher who can successfully guide the ‘child’ through the eating/learning process. AI demands a human in the loop.

It is unfortunate that during the past 60 years we have not learnt more and uncovered and used further aspects of the human brain. Isn’t it about time that we stopped and looked at current projects, and tried to learn more from them? True, there has been good progress in mathematics, but what about the lessons from biology?

There is now more understanding of synaptic shrinkage but almost nothing on synaptic expansion, the reason the surface area of the brain appears to be maximised, how we could use the cerebellum for macro-like commands, or how we can use the brain’s menu and information guidance structure, the medulla oblongata.

How long will it take us to learn from the best computer on the planet, our own brain?

Andrew WS Ainger CEng FIET

By email

Pragmatic solutions for developing countries

I would have to question the supposed pragmatic approach to energy generation in developing countries proposed by Professor Phil Hart (Comment, January 2022). In his analogy, he suggests that forcing developing countries to stop extracting fossil fuel and instead buy an expensive hydrogen stove for £250 is not practical. Instead, he proposes allowing them to continue extracting fossil fuels to help fund purchasing the stove at some unknown future date – by which time it will have probably been superseded by an even more expensive stove, and carbon-emission targets will have been missed.

This is driven by the same economics that led to the environmental issues. A more pragmatic approach would be for developed countries to honour their pledges to fund and facilitate technology transfer. The resultant increase in productivity, supported by writing off debts and reducing trade barriers, would reduce carbon emissions while supporting economic development.

Rainer Hurricks MIET

By email

Who benefits from smart meters?

John Cowburn’s letter (E&T April 2022) on the challenges of smart meter requirements prompts me to question their raison d’etre. Once upon a time I had a normal meter. Three times a year I would get a card inviting me to enter the numbers shown on a large screen and post it. Once a year an affable person would call to check the numbers and see if all was well. Simple.

For the last few years I have been blessed with a smart meter. Now I get an email every month asking me to sign into my account and key in numbers, which can only be found in tiny font at the bottom of the meter after pressing one of six fiddly little keys six times. I have about as much chance of getting this ritual right first time as I have of picking a Derby winner.

Richard Bristowe MIET

By email

Not as simple as changing a light bulb

The amount of energy consumed by ‘vampire’ electronics that are left on standby has been in the news recently. I have a plug-in electronic W/V/A/Wh meter that gives credible readings in most domestic situations, but didn’t with a 20W LED tube in an old fluorescent fitting. I found on replacing the old tube that the power consumption was still apparently 55W, not the expected little over 20W.

Could it be that the meter is confused by the peaky current waveform, or does the waveform cause significant losses in the ballast inductor that would still be in series with the LED tube? Removing the ballast cut the power down to 22W. Has anyone had similar experiences? If so, should there be a warning that simply putting a new LED tube in an old fitting may not be as economical as expected?

Geoff Tily

By email

The myth of degree apprenticeships

Having dispelled the myth that the only path to a worthwhile career in engineering is via a full-time degree course, the Back Story article in the April 2022 issue of E&T then unfortunately perpetuates another: that degree apprenticeships are new.

Those of us of a certain age will have been well aware that what were then called sandwich courses were readily available, and equipped graduate engineers with a range of on-the-job practical experiences in addition to their academic qualifications. I graduated in 1973 from Brunel University with a degree in electrical and electronic engineering after four academic years in simultaneous employment as a student apprentice with a major electrical and electronic equipment company.

I know that since the turn of the century, apprenticeships in general have been touted as the big new idea, and thank goodness we are realising their value again. But new? I don’t think so.

CJ Poole MIET

By email

UK marine energy’s delayed start

I support much of Leon Freris’s letter (E&T March 2022) about the UK’s ‘relentless obsession’ with nuclear energy, but would like to draw attention to one omission in his advocacy.

In the 1980s, I project-managed for Ewbank Preece Energy Projects a study of the UK wave energy programme for the Department of Energy/AERE in association with Gifford & Partners. In our report’s concluding recommendations, I pleaded with the government to resist its policy of reducing energy prices as North Sea gas and oil came ashore, and instead fund marine energy R&D.

That administration and its successors ignored the pleas and we lost 30 years before the European Marine Energy Centre was launched in the Orkney Islands. Marine energy – a predicted 20 per cent at least of UK total demand – could easily displace dependency upon Russian gas and allow us to sell our very considerable expertise abroad. We should long ago have taken control of energy out of the hands of short-term-thinking politicians and made it the responsibility of a long-term-thinking committee of independent professional advisers and stakeholder.

Brian Mallalieu MIET

Bakewell

The real reason why ships won’t go nuclear

With reference to the article on the use of nuclear power for commercial shipping (E&T March 2022), in 2014, I was asked by my project manager to produce a feasibility study on the application of PWR technology, as used in nuclear submarines, to commercial shipping. The obvious candidates were cruise ships and very large cargo ships carrying oil, or solids like iron ore. I concluded that the spent nuclear fuel should be capable of being reprocessed at civilian facilities, so it would need to be similar to that used in nuclear power stations. If the design could be modular, this would mean modules could be swapped out, making refuelling easier.

At that time no nuclear-powered ship could get insurance, because the risk was impossible to define, and my study concluded that the idea was not feasible.

In the summer of 2019, while on my retirement cruise, I asked the ship’s captain if he thought cruise ships would ever use nuclear power. The captain, and his colleague who was acting as host for a Q&A session, both stated that the future was LNG. The captain’s final remark on the topic was “Who would want to float around the sea, sitting on top of a bomb?”

Even if the technology is proven to be safe, with the risks quantifiable and insurance available, convincing crews and passengers that it is safe might be the more difficult challenge.

David Watson MIET

By email

Magnox outside the UK

Keith Sylvan asks (Letters, March 2022) about Magnox nuclear power stations outside the United Kingdom. The UKAEA was required to create an atom bomb and built the Windscale facility to produce the uranium it needed for the bomb and for power generation. It was decided that the UK should require the CEGB to commission five Magnox power stations.

To design and build these, the UK created five consortia who would each design a different power plant design so that performance could be compared. The consortia, having designs that had been accepted by the CEGB and construction commenced, then began to market their designs throughout the world, with the UKAEA supplying the fuel elements. Two overseas contracts were won: Latina in Italy to ENI and Tokai Mura in Japan to Japanese Electric Supply.

I was responsible for the supply of all the fuel elements to the UK, Italy and Japan. This is a fantastic story, but almost none of it is known.

FL (Mick) Pitt

By email

The best way to save energy

Derek Rimington is correct – the ‘energy-saving’ devices he has seen advertised (Letters, March 2022) are intended to provide a form of power factor correction in a domestic environment but, as explained by an excellent debunking on the Electrical Safety First website (www.electricalsafetyfirst.org.uk), are ineffective.

Apart from finding a cheaper energy supplier (good luck with that) the best way to save money is to turn something off!

Paul Allen FIET

Longfield

Write to E&T at engtechletters@theiet.org. We reserve the right to edit letters and to use submissions in any other format.