Computing checks, AI, water and datacentres – Laptop Weekly Downtime Upload podcast
Hear to this podcast
On this episode of the CW Downtime Upload, Clare McDonald, Brian McKenna and Caroline Donnelly discuss the most modern A-level and GCSE computing results, AI appropriacy and datacentre water consumption
On this episode of the Laptop Weekly Downtime Upload podcast, Clare McDonald, Brian McKenna and Caroline Donnelly discuss the 2021 A-level and GCSE computing results, what algorithms are and are now not gorgeous for, and the water consumption habits of datacentres, and their environmental impact.
Clare opens up the episode with an fable of the most most modern A-level and GCSE results in England, Wales and Northern Ireland – a bumper carve of examination results which are usually a small extra spaced out.
This year’s results safe been algorithm-free, unlike, in the initiating, the 2020 results. But now not freed from their celebrated class bias: non-public college college students did even better than celebrated.
Clare notes the tense Covid context for this year’s college students. No matter that, there used to be an even bigger-executed suppose of A-level computing results. There used to be a upward push in the series of college students taking A-level computing: 13,829 college students in the UK took computing at A-level, a upward push from 12,428 entrants the earlier year.
There used to be a year-on-year lift in the series of women taking A-level computing nevertheless additionally, worryingly, a fall in the numbers taking GCSE. On the assorted hand, girls are doing better. For women, 25.7% executed an Aresult, a upward push from 17.8% final year; whereas simplest 18.9% of boys executed an Alevel grade, a upward push from 13.1% final year. For the first time, girls additionally outperformed boys in arithmetic in A-ranges and GCSEs.
Algos – what are they gorgeous for?
The episode then moves on to a related discussion about what algorithms and artificial intelligence (AI) are gorgeous for, if they’re now not gorgeous for checks.
Brian touches on the A-level and Scottish Highers debacle of 2021, when algorithms obtained themselves a mistaken establish. This used to be talked about on the podcast practically exactly a year in the past.
The whole fiasco used to be talked about in a extra most modern BCS file, Priorities for the nationwide AI technique, written by Invoice Mitchell, the BCS’s director of coverage.
The file says the UK should buy an world lead in AI ethics if it cultivates a extra various crew, including folks from non-STEM backgrounds. It refers help to the Ofqual algorithm that used to be earlier to estimate GCSE and A-level grades in 2020. In the BCS creator’s words, this led to a “neatly-liked public mistrust in algorithms making excessive stakes choices about folks”.
The file registers that “public belief in AI and algorithmic programs in celebrated has been seriously eroded by occasions one day of the pandemic”, as shown in two nationwide surveys by YouGov commissioned in 2020 by BCS. These chanced on that:
- Over half (53%) of UK adults wouldn’t safe any religion in any organisation to exercise algorithms when making judgements about them, in points ranging from training to welfare choices.
- 63% of UK adults disagree with the statement “Students graduating with a computer science college degree are qualified to put in writing application that makes existence choices about folks”.
The BCS file itself will feed into the executive’s AI technique, which may per chance per chance additionally also be published later this year.
On the podcast, Brian says the headline in it for him used to be the premise that for the public to belief AI programs, we desire a broader suppose of folks making them in the first location. Now not that it’s all bleak for algorithms. Where would we be without the recommendation algorithms of the streaming services we have trusted so grand over the pandemic?
Brian then poses the ask: are there areas of human existence that ought to gorgeous be freed from algorithms?
Caroline cites one rental that ought to be off-limits to algorithms and data analytics: Queen of Pop Beyoncé’s career, as illustrated in an interview in Harper’s Bazaar. Beyoncé’s refusal of data analytics-based fully fully recommendation referring to the 2008 album I am…Sasha Fierce is an exemplar of retaining with “the human feeling and spirit and emotion in my decision-making”, she says.
The team discuss some various areas where AI will be mistaken, equivalent to recruiting, nevertheless additionally will be purposeful. We may per chance per chance additionally, despite the big hype, be in the early days of how AI will transform human lives. That is, if folks continue to exist climate catastrophe.
Water and datacentres: a matter below-talked about in climate swap debates
In the third share of the episode, Caroline discloses some work development on the topic of the water consumption habits of datacentres, and what these may per chance per chance additionally indicate for the atmosphere in the context of the increasing climate disaster.
She reveals how her consideration used to be in the initiating drawn to this theme by a observation at a commerce conceal in 2016 – that it’s a long way never known how grand water is earlier by datacentre operators in their cooling programs.
Datacentres safe advance below scrutiny for his or her environmental impact and sustainability efforts alongside three dimensions, explains Caroline: how grand energy datacentres exercise, how grand of that energy is renewable, and the draw good the carbon footprint of datacentres is. Development has been made on these three fronts.
There may per chance be a broadly earlier enterprise metric for measuring the energy consumption of datacentres, PUE – the energy utilization effectiveness rating.
And while datacentre operators are appealing about disclosing their PUEs, they’re oddly quiet about one more measure – their water utilization effectiveness. There may per chance be a metric, published in 2010, that the datacentre enterprise may per chance per chance additionally exercise for the water utilization efficiency of their datacentres – WUE (water utilization effectiveness). But it certainly is either now not being earlier, or the scores are now not being published. Fb is a partial exception, nevertheless that’s of small moment to enterprise IT consumers.
Caroline cites an Uptime Institute survey from 2020 that acknowledged half of all datacentre operators make now not measure how grand water they exercise. Would per chance presumably also this be because operators are the exercise of hundreds of water to deflate their PUE scores?
Water is a precious handy resource, in particular so in areas which would be an increasing number of self-discipline to drought, equivalent to California, or which suffer a celebrated stress on the provision of drinking water. These embody countries equivalent to Spain and Singapore. And that roster of water-harassed out areas may per chance per chance additionally salvage bigger in years yet to advance – datacentres are long-time duration parts of the landscape.
Novel air cooling will be segment of the technique to the over-exercise of water. One other may per chance be the datacentre-on-a-barge advance being pioneered by Nautilus Records.
It is a long way early days for this topic, says Caroline: “Water is one thing that datacentre operators are now not speaking about, nevertheless with climate swap, and the possibility of that changing into grand extra obvious and proper than it beforehand used to be, that can should swap. And there needs to be extra tension on operators to be extra clear about how what they make impacts water supplies.”