My thing about heat pumps is that there’s a big gap in the policy discussion of this: the hundreds of thousands of properties which are houses converted into 2-8 flats, with a mix of tenure.

They’d be best served by one heat pump for the whole building, but it’s going to be hellish to organise that

I am a highly educated and technically proficient person, but passkeys are the first new ubiquitous web technology that regularly make me think I am just a stupid big baby man

Where is twcuddleston these days anyway?

Tried a Boba Tea when Heytea opened on High Holborn and…I don’t think it’s for me.

Journalists: the soon-to-open Obama Center is *not*

the first digital-only presidential library

a presidential library at all

a federal archival depository

funded by federal taxpayers

funded only by private donations

Brutalist architecture

and it will not be owned by the Obama Foundation

Look, I know that Palantir products are apparently having good results within the NHS (see recent articles in the FT) but I don’t think it’s unduly fussy to question whether the British state should be dealing with companies like this.

Having, a long time ago, worked at NEST (and its predecessor PADA which set the thing up) I think that’s unfair about them at least.

Not sure if Mark Fawcett is still there but ehe did a really good job of looking at the investment side from first principles in a really thoughtful way.

I switched to Starling recently and the way that all the challenger banks have easy links to crypto wallets is disturbing.

Ian Fairbairn, who founded it in memory of his wife, was a key figure (at M&G) in bringing Unit Trusts to the UK.

He was a huge believer in the benefits of diversified equity investments for all and, aiui, that was one of the reasons he set up the charity.

It got very rich from his M&G shares.

I do find it a little odd that the Esmée Fairbairn Charitable Trust, which has one of the largest endowments in the UK (and which, to be clear, does great work) appears to have so evolved its goals and objects that it no longer does anything on financial education. /1

My feeling is that for a lot of use cases they’re pushing the wrong product.

The combination of semantic search and retrieval augmented generation is, I think, far more useful in a lot of fields. The likes of Copilot can’t do it but other products by the same companies can.

‘Such a mix of people’: Ireland of 1926 was not monocultural, release of census shows Archive is freely available online from 18 April, revealing the lives, occupations and secrets of 2.9m people

The 1926 Irish National Census has been digitised and is being made available online from today - a treasure trove for researchers and for families!
www.theguardian.com/world/2026/a...

When I did Paris Marathon in 2012 (my second) I got caught out by the fact that all their fluid stations were water only apart from one at Mile 22ish where you 1/3 of a plastic cup of Powerade.

My own fault for lack of research but it was a nasty surprise.

locations was practically impossible thereafter for large numbers of men deployed in about 120 mobile columns, tracking down Boer guerrilla forces across a vast area. This was not a style of warfare conducive to postal efficiency, but the verdict at the end of hostilities in far-off South Africa was nevertheless that the Army Post Office Corps had more than proved its mettle. Given a combat zone with clearer front lines and shorter lines of communication, the capacity of the Post Office and the War Office together to bring mails to the brink of the battlefield could hardly now be doubted. The partnership arrangement was restructured in the years that followed. A fundamental reorganization of the army in the years after 1906 ('the Haldane reforms') included the setting up of a Territorial 'Army Postal Service'. After some years of confusion, this emerged as REPS.* Responsibility for REPS was handed to a Territorial officer, tasked with organizing a mobilization plan: the postmen reserves, in Royal Engineer uniforms and subject to full military discipline, would need assigning to all divisions of the regular army in the event of war. (The 24th Middlesex was meanwhile transformed into the Post Office Rifles, which remained quite separate.) The man appointed as the first 'Director of Army Postal Services' was William Price, who was given a full establishment of 300 officers and men - enough, it was reckoned in 1913, for the six divisions of any Expeditionary Force likely to be sent overseas. Price himself would be attached to the Army's Inspector-General of Communications and would report to the Quartermaster General. Here at last was a structure that looked equal to the demands of the hybrid concept. By the time those first REPS men set off with the BEF in August 1914, confidence in its effectiveness was high.! Two nasty shocks followed, much as they had done in South Africa - and identical in nature. The scale of the action had been badly underestimated, quickly leaving Price with far too few men. This problem, though, was soon redressed: several hundred more were recruited in haste from the ranks of the POR battalion (a return, ironically, to the model left behind with the 24th Middlesex), so trebling the size of the REPS force in France. The second early crisis sprang from the speed of the enemy's advance in the opening weeks of the war. The Germans' push for Paris engulfed several of the first postal facilities, established in and around Amiens; no sooner had others been set up than they had to be hurriedly abandoned, and thousands of bags of mail were hastily shuttled back from the front lines by train, lorry and rented boats ever further to the west, ending in near chaos at the port of Nantes on France's west coast.13 Once the Germans' Schlieffen Plan had failed to win an abrupt victory, a very different prospect opened up. The static warfare of the trenches - whatever its horrors in every other respect - posed the kind of logistical challenges that were meat and drink to the men from the Post Office. The BEF's layout, while embracing ever more army divisions, could in effect be treated as stationary from the Channel to the front line. Of course, individual units were still shuttled regularly between different locations, but most of those locations themselves remained as fixed as London suburbs. After some initial months of trial and error, REPS built a delivery system to match. By the onset of the winter in 1914, it had raised its force from 900 to 1,500 to cope with a huge increase in the mails for Christmas. The result, amid the grimness of the Western Front, often verged on the bizarre. Letters were usually reaching addressees in the trenches on the second day after being posted, and London newspapers (also sent over in great numbers) were being widely distributed within a day or two of their publication. As one officer confided to his diary on 30 December 1914: 'I got three letters posted in Ireland on the 26th and in England on the 28th, this afternoon, so letters are reaching us as quickly as if there were no war. It's really rather wonderful. "4

2. POSTAL SUBURBIA ON THE WESTERN FRONT To say that postal services on the Western Front marked the high-water mark of the Victorian Post Office might be stretching a point - but not by much. The REPS contingents working with the BEF, though they grew steadily through 1914-15, never numbered in total more than a few thousand men. (The numbers recruited into the Signals Service, and later the Royal Flying Corps, were much greater from the start: telephones on the Western Front relied on the work of some 19,000 men from the GPO.'5) Yet this modest postal force was able to impose order on a vast network of interwoven roads and railways that linked London and the Channel ports with up to five army groups, each of them roughly representing the needs of a provincial city the size of Southampton. Losses of mail due to accidents or enemy action were extremely rare. Of course, the task demanded courage and initiative in the ranks far beyond anything normally expected of peacetime sorters and postmen. But it was also achieved on the back of traditional strengths that made the GPO a natural partner for the Army - harnessing the meticulous attention to detail, the strong internal disciplines and the by-the-book culture that were all so central to the Victorian postal legacy. Plainly, outward mail could not be sent directly to a serving soldier or sailor using any normal address. The Post Office issued a series of notices advising the public on how to address letters to men at the front. Only the details of the addressee's fighting unit could be used, and all letters for the British Expeditionary Force were then diverted by the Post Office to a clearing centre where they could be sorted for onward despatch. This was 'Home Depot', an operation that came eventually to employ 2,500 staff. It also handled inward mail from serving men and trained around 5,500 recruits, who went off to join the swelling ranks of REPS men serving in France and other overseas postings. The Depot was under the nominal command of the Army and employed many uniformed men, but it was controlled by the GPO: it was run by the pre-war postmaster of the capital's WC district, now installed with the pay and allowances of a REPS captain.! It began life scattered among different buildings, but by the end of 1915 had been pulled together in a vast hut erected across five acres of Regent's Park, reputedly the largest wooden edifice in the world. A fleet of 220 three-ton army lorries was eventually kept busy ferrying mailbags to and from the trains which linked London to Folkestone and Southampton. Several steamers a day left these ports for 'Base Post Offices' in Boulogne/Calais and Le Havre respectively, where formal responsibility for the mails passed from the GPO to the British army. Most outward bags were loaded onto trains that plied back and forth across Picardy under cover of darkness, some of them to be dropped at stationary 'Army Post Offices' along the way, linked to the Lines of Communication, or at large troop placements (such as labour battalions or artillery placements) that were a semi-permanent fixture. The remainder went on as far as the railheads that serviced 'supply trains' to individual divisions; every railhead eventually had its own post office, running a branch network between the railhead and the trenches. The mails usually left the railhead in special REPS lorries, which took them to the spots designated for the collection of all divisional supplies, known as 'refilling points'. Here they were taken up by the Brigade HQs. The final stage of their journey was described in an article in the St Martin's-le-Grand Magazine of July 1915 by 'F.H.W. - the newly appointed head of REPS on the Home Front, Frederic Williamson, just back from an initial tour of the lines to see for himself how things worked:

At some pre-arranged point, the regimental post orderlies drive down [to meet the motor lorries from their Brigade HQI to receive their mails. The mails are stacked and sorted by the roadside; the long line of carts comes up, each in turn to receive its quota of the mails from home; and each orderly then clatters off with his letters and parcels, which he carries to his comrades right up to the firing line.19 It was normal practice in the trenches for each day's post to be handed out with the evening meal by ration parties. These would also collect the men's letters and postcards for home, which - unlike in the Boer War - could be posted without charge. They were generally first collected together at 'Field Post Offices' - large black steel boxes, usually sporting a pole with a distinctive red and white striped flag, which were relocated from time to time as the need arose. Most FPOs had all the kit to provide a good range of the usual counter services available from a village post office in England, from the sending of registered letters to the buying and cashing of postal orders; after the summer of 1915, they even sold war savings certificates and war loans. Letters were stamped with a field postmark, from January 1915, showing the date and an APO or FPO number (and making rich pickings for collectors in decades to come). Then the postmarked letters were sent up the line via the Divisional Railhead Post Office and Base Post Office for the return journey to Home Depot.2° Four years after the war was over, the British Army made a thorough study of the Field Postal Service that had been used by the Kaiser's army. It had been substantially bigger, and had delivered about four times as many bags of mail each month. The resulting 200-page report nonetheless compared it unfavourably with REPS on many counts. It concluded without any reservation at all that there was little of value to learn from the German experience'. " For all their prowess in many dimensions of military planning, and despite their experience of three nineteenth-century wars of their own, the Germans seemed to the authors of the report to have spent the second half of the war constantly reorganizing their posts and scrambling to copy a British model that had endured to the end and proved significantly more efficient. It was a proud boast - but not an empty one. Of course, the division of responsibilities between REPS and the Post Office did evolve over time, and developed in some ways that were not remotely foreseen at the outbreak of the war. The Postmaster General, for example, agreed to answer in Parliament for the whole operation. This had not been anticipated, but seemed sensible given that the GPO not only provided many of its personnel but also made all of the transport arrangements as far as overseas Base Post Offices." In general, though, the structure laid down by the end of 1915 needed few substantial changes. And through the remainder of the war on the Western Front - leaving aside the temporary crises precipitated by the Germans' spring offensive in 1918 and the final Allied counter-attack - it expanded steadily without serious mishap. The critical importance of letters to the morale of the troops was acknowledged from the outset: as everyone agreed, letters came close behind food, leave and suitable clothes, and the press wasted no opportunity to remind readers of this truth. At an early stage in the war, one of the London papers published a letter from a soldier in the trenches saying he was lonely and would appreciate a letter. In the ensuing weeks, he reportedly received 3,000 letters, ninety-eight large parcels and three mailbags of smaller ones? But even the Post Office was taken aback by the vertiginous growth of the three main categories of outward-bound post to the Western Front.4 (See Appendix A, Chart 4.) The Christmas peaks were scarcely less impressive than the growth from one year to the next: not far short of 5 million parcels went to France in the month before Christmas, 1916. Supplies of tobacco and cigarettes - given duty-free status by the French from the start - remained a staple item at all times of the year. (Matches to light them with were a more complicated matter.

Faced with a dire shortage of them in the trenches, the War Office in October 1914 prevailed on postal officials to lift a ban on matches that had been in place since 1884. But it was waived against the better judgement of the Post Office, which specified elaborate rules on how safety matches (only) were to be packed for postage. Several serious fires in mail depots in 1915 were nonetheless blamed on the ignition of matches, and the ban was reimposed in May 1916. It was diligently enforced, too: the Post Office brought 178 prosecutions against offenders before the end of the war.2) Outbound parcels soared to just over a million a week by the spring of 1917, though food shortages at home, compulsory rationing after December 1917 and, eventually, much improved field canteens at the front saw this number contract steadily (except at Christmas) through the rest of the war. Sending cash was a much better option, and accounted for a high volume of registered letters in the last two years of the war. Outbound letters peaked at more than 12 million a week early in the first quarter of 1918. (In the last year before the war, by way of comparison, about 66 million letters and 2.5 million parcels were posted in Britain during an average week.) Behind these bare numbers were mountainous piles of mail that needed shifting at each stage of the journey from Home Depot - entailing, at many stationary Army Post Offices, the arrival of 400-500 mailbags every single day. While the Western Front accounted for more than 90 per cent of the numbers, it was far from the only theatre served by Home Depot. By the time the War Office was ready to launch another Expeditionary Force, early in 1915, sending any significant number of troops into action without an Army Postal Service unit in attendance was out of the question. Wherever combat forces were to be engaged, REPS men would have to follow - and usually not far behind, as a post-war collection of operational reviews recounted.26 These were never published, but offer a useful commentary on many of the campaigns beyond Flanders. In the Dardanelles, REPS men sorted mail in trenches along the beach within range of Turkish snipers. Through that disastrous campaign, the unopened letters coming back down the line were almost as numerous as those going up it, so heavy were the casualties.* By the time five shattered divisions from Gallipoli were evacuated to Salonika that autumn, two hundred or so peacetime postmen were running a Base Post Office and five FPOs in the still sweltering heat - and most of them remained in the Balkans for the rest of the war, as the post-war review put it, 'buoyed up by their sense of duty as soldiers and their loyalty to the Post Office'. Other units of the defunct Mediterranean Expeditionary Force went east via Sinai to Syria and west to Italy. The REPS network in Italy covered a range of 1,000 miles from Tarranto to the Alps and reckoned to deliver all letters from GHQ to the front within twenty-four hours. ('This, it may be remarked, was considerably quicker than the Italian Civil Post and infinitely more reliable') Perhaps most remarkable of all was the assignment of a REPS detachment to Murmansk and Archangel in the autumn of 1918. Its men had to contend not just with the Arctic winter but also with inter-service squabbles at home, where the Sea Lords of the Admiralty 'were instructing the GPO when to send out mails for the Navy in North Russia but ignoring the existence of the Army ...' In the process of servicing the Allied troops sent to intervene in the Russian Civil War, these hardy souls found themselves handling around 2,000 letters a day for the White forces. By the summer of 1919, the head of the unit in Murmansk was writing home to William Price in optimistic vein. 'Should ... the Bolsheviks [be] cleared out of Petrograd, the question of sending mails for North Russia via the Baltic to Petrograd will have to be considered?

From ‘Masters of the Post: the Authorized History of the Royal Mail’ which I happen to be reading for work.

Aww, congratumalations!

The bits in Andor where they showed the Fox-esque coverage of Ghorman were pretty pointed.

But you don't need to train the model on this material.

I'm talking about using RAG and/or vector indexing etc. I admit I'm moving out of my area of technical expertise here, but aiui there are ways to combine the pre-trained LLM with techniques to index unstructured trusted material.

I'm not sure I agree?

In specialist environments (e.g. within a business, an academic department, a GLAM institution) you are quite likely to have access to a big, trustworthy, document set dealing with a specific subject.

As Copilot itself told me, it "retrieves and summarises; it does not truly index and reason".

For a lot of use cases, I think the wrong AI tools are being promoted.

A lot of people/companies need semantic search - to index a large corpus of trusted material so you can easily find reliable answers.

The likes of Copilot can't do that - but tools like Azure AI Search and AWS Kendra can.

Have you ever heard Stephen Fry imitate his Austrian Jewish grandfather's way of pronouncing it?

Very funny and sweet.

I honestly can't remember which episode, and indeed whether it was on the main 'Kermode and Mayo's Take' or the subscriber-only 'Take 2'. It was probably about 4 or 5 weeks ago.

Wonderful to see a really artful/ useful application of AI tools to a news archive. (Also super cool that it started in @columbiajournalism.bsky.social Lede program …h/t @dangerscarf.bsky.social )

He spoke entertainingly witheringly recently on the Kermode & Mayo podcast about Tarantino's stupid remarks on Paul Dano and Matthew Lillard.

The annoying thing is that, as I understand it, there *are* AI tools that will do this kind of thing - and do it very well.

But for whatever reason all the promotion is of generative AI tools that won’t.

It’s a weird one. Bain’s invention predates the telephone by 40 years but it was never viable as a consumer-facing product.

Likewise picture telegraphy was around by about 1900 (it’s how foreign photos were printed in newspapers) but the tech was big
and costly.

Cheap fax machines arrived c1980.

There is definitely something in this.

And aside from celebrities and other rich people, even at the very mundane level, lots of people post things (especially around the self-assessment deadline) that make it clear they’re higher rate taxpayers and they sort of assume everyone else here is too

Faxed The Rise and Fall of the Fax Machine

I have this on my e-reader waiting for me to get to it. press.jhu.edu/books/title/...

The fact that Alexander Bain patented an electro-chemical printing telegraph in 1846 is neither here nor there.

The key technical developments were in the 1960s and the fax only started to gain mass adoption by business from 1980 after the agreement of interoperability standards and cheaper models.

Also the 180 years thing is misleading. Alexander Bain’s chemical printing telegraph was never commercially viable.

The fax as a mass market technology is 40 years old.