Re: Free pass
If the conditions of buying the book include that it should not be stored in an electronic retrieval system then copying it into such a system is actionable by definition - unless the condition is, for some reasonable, unenforceable.
40471 publicly visible posts • joined 16 Jun 2014
"SOFCs are more efficient than gas turbine generators, and the transmisson loss from on-site generators will be minimal compared to power coming from the grid."
Where does the hydrogen come from? If it's from electrolysis you have to consider the transmission losses to the electrolysis plant and the electrolysis itself. Where is the electrolysis plant? if it isn't on-site you have to factor in the energy needed to pump it and, hydrogen being hydrogen, the losses from the joints in the pipework, the replacement of the pipework due to embrittlement etc. If the plant is on-site the transmission losses are the same as you'd have had powering the data centre direct plus an addition to the second order effect of transmission losses incurred in transmitting the energy that's lost in the course of hydrolysis.
If a fuel cell facility helps stabilize the grid then surely this is a matter for the grid operator rather than a grid customer.
"powered initially by gas, moving to hydrogen in future."
Leaving aside the fact that hydrogen is also a gas, powering the fuel cell by natural gas, i.e. methane, is hardly decarbonising the operation compared to using the same gas to fuel a gas turbine-driven generator. The significant question is how is the long-term hydrogen to be obtained? If that involves simply transferring electrical energy from some other source into chemical energy in the form of H2 why not use that electrical energy directly?
Fuel cells and hydrogen make sense (give or take the difficulties of handling hydrogen) where the energy is to be deployed in situations which are intrinsically disconnected from the grid such as vehicles. The only point I can see for a static installation such as a data centre is green-washing.
It's not necessarily lack of staff training that's the problem. The staff training might have included extracting data into a spreadsheet. The underlying problem is more likely to be lack of a proper procedure as in Mike 137's post and insistence on a format that precludes any hidden content that might escape initial inspection.
Our statistics class was more or less similar. We handled it be turning up a bit later each week so he also turned up a bit later. Then we turned up on time and he wasn't there so we left for good.
Doing research the best way to use statistics was to call on a statistician for advice. Rather like encryption, if you're not a specialist don't try to roll your own.
"and then bang through loads of equations on the board at speed usually missing out a few steps"
This was my experience of (an admittedly small number of) maths teachers. I think, at least in my day, it was characteristic of many teachers and probably is of many of us - including IT support.
Different disciplines require different modes of thinking. It is very difficult to put oneself in the position of someone who doesn't think in that mode. For non-mathematicians following a batch of equations takes time, especially if it involves wrapping the mind around a newly introduced notation, To the mathematician what's quite clear looks like sleight of hand to the rest of us. Not realising this the maths teacher might spend more time than necessary on initial assumptions and still leave the non-mathematician none the wiser when the dust has settled.
Much the same thing applies to parsing a highly inflected language such as Latin, at least for those brought up to speak English.
"Lastly, human beings have fallen for the likes of Musk and Bank-Friedman for millennia."
Hutton's next sentence is unfortunate:
Instead of uncritically lionising all things technological and mathematical, never forget the lessons of undervalued literature and history – the best antidotes to a confused and credulous present.
The best antidote is sufficient technological and mathematical knowledge to see through charlatans and their snake oil. I suspect those who fall hardest for them are those best versed in literature and history.
"There is no domain of knowledges being formed with propositions, tests and experimentation, just the munging together of scraped content from elsewhere which may, or may not, be accurate and is likely rather biased too."
If the models are recursively trained on the output of models what is the long-term outcome? Do they converge so they all give the same, possibly meaningless, output to every prompt or do they diverge and emit random strings of nonsense?
"So maybe messages were held back 48hrs, or that was how long it took messages to get from collection, filtration and to the relevant LEAs."
Apparently the messages were sent as an overnight download to UK. NCA then filtered it into separate batches for the various units who would have to deal with it. A message intercepted just after the previous download would would be ~24 hours old by the time of the next download so allow some processing time for sorting and her unit to upload their batch onto their own system & 48 hours sounds about right. But depending on the timing of the intercept vs the overnight download schedule it could have been a good bit shorter.
There's an account elsewhere but it would make a good el Reg article.