dtf

These release notes are very interesting! I spent a couple of weeks recently writing a ProRes decoder using WebGPU compute shaders, and it runs plenty fast enough (although I suspect Apple has some special hardware they make use of for their implementation). I can imagine this path also working well for the new Android APV codec, if it ever becomes popular.

The ProRes bitstream spec was given to SMPTE [1], but I never managed to find any information on ProRes RAW, so it's exciting to see software and compute implementations here. Has this been reverse-engineered by the FFMPEG wizards? At first glance of the code, it does look fairly similar to the regular ProRes.

[1] https://pub.smpte.org/doc/rdd36/20220909-pub/rdd36-2022.pdf

This codec needs hardware support to be practical (ie recording live video rather than just transcoding). So far, I think that's only available on certain Samsung models.

I wonder if these new devices will finally have support for the elusive APV codec.

https://en.m.wikipedia.org/wiki/Advanced_Professional_Video

You'll need more than just an account to access "certain mature content" on sites like Reddit - you'll soon need to upload some photographic ID.

https://www.bbc.co.uk/news/articles/cj4ep1znk4zo

The diplomatically neutral term is "the island of Ireland" for the whole thing.

"These islands" refers to the British Isles, in the context of Anglo-Irish relations.

It's some fascinating wordplay, born out of the Good Friday Agreement, trying not to step on various people's buttons.

Deno KV uses FoundationDB, for example:

https://deno.com/blog/building-deno-kv

I just googled a bit and found his race report from that event, dated April 1995:

https://www.mattmahoney.net/barkley/bark100.txt

35

As far as I remember, it's down to security concerns.

You can actually insert your HTML into a SVG foreignObject, and then drawImage() that onto your canvas. But your HTML-in-SVG document will need to load all its own resources, fonts, CSS etc.. which makes this process rather tedious.

What on earth happened to this once powerhouse of engineering and innovation?

Also worth mentioning Manuka, which is Weta's production spectral renderer. There's a nice overview of the architecture here, which includes a section on their motivations for, and implementation of, spectral light transport:

https://jo.dreggn.org/home/2018_manuka.pdf

There was a piece in the Guardian this morning arguing that it differed by country, and in particular that in countries with vocally anti-migrant governments such as Italy, it had become less of an issue.

https://www.theguardian.com/commentisfree/2024/jan/22/crisis...

"Germany is the only country where immigration is clearly in the lead when people are asked which crisis bothers them most when they think about the future. Estonians and Poles are focused on the war in Ukraine. Italy and Portugal see the economic crisis as their biggest threat. Spain, Britain and Romania are the countries where Covid-19 is seen as the biggest trauma. And in France and Denmark the climate emergency is considered the most important crisis."

Why do people pay premium for "hand made", "artisanal" etc?

If AI is cheap, then by using AI generated content to cut costs, you will cheapen your brand by association. There are examples of brands doing this to themselves already.

Some bright spark at CCHQ realised it could be weaponised against both Ed Davey (Leader of the Liberal Democrats, and former Post Office minister 2010-2012) and Keir Starmer (Leader of the Labour Party, and former Director of Public Prosecutions 2008-2013).

I was kind of shocked to find this out too. For instance:

"As of 2020, 78.7% of Singapore residents live in public housing, down from a high of 88.0% in 2000."

https://en.wikipedia.org/wiki/Public_housing_in_Singapore

The risk premium was access to a large and very cheap credit facility:

https://twitter.com/jonwu_/status/1634250770555219970

First, there's really nothing small cost about Bannister's stopwatch:

https://www.alamy.com/stock-photo-the-omega-bannister-stopwa...

Sure, I can pick up the awesome Casio F-91W today for £10 and it would do much the same job. But for the modern day money equivalent of that gorgeous Omega piece I'm going to be able to afford a high-end GPS watch. There's nothing comparatively exponential about it.

Accuracy requirements will depend on training context - if you're running 60 second laps on a track and trying to shave 1 second of your mile PR like Bannister was, then you'll be wanting a certified track and decisecond accurate clock. If you're training for the marathon and running 5K repeats on the road, a few 10s of meters or seconds here or there doesn't really matter.

I can understand why some people don't want to run with a smartwatch, measure themselves, broadcast their progress to all on social media, judge themselves against others, and so on. (I also don't generally judge those who do, unless they're truly awful!)

I can also understand why some people would rather not run with any measure of time or distance at all. They feel it gets in the way, they'd rather just run free, they'd rather just run for fun, or they'd rather just race others for places in the spirit of pure competition.

I just can't get my head around the concept (and this isn't the first time I've heard it), that somehow old-school watches are acceptable but modern watches are bad. It smacks of neo-luddism (or sometimes hipsterism).

Bannister's use of a mechanical stopwatch on a 440 yard track is as much a case of cybernetics as a modern GPS or a heart rate monitor. He obviously didn't just turn up on the day and run 3:59.4. He trained for months with a stopwatch:

"Several days [per week] consisted of 10x440 in 66 seconds with a 2 minute rest. During the following months they were gradually speeded up ... to 59 seconds per 440." [1].

So aside from accuracy, what's the difference between training with the feedback of timed laps on a track (be it Bannister's mechanical stopwatch and cinder track measured in imperial units), or a modern athlete running kilometer repeats on the road via their fancy Apple or Garmin smartwatch?

If one is that much of a running purist, why measure time or distance at all?

[1] https://twitter.com/jmarpdx/status/1465431668206944256

I've had that sometimes during the cold seasons, and usually take a puff of salbutamol before setting out. I still get it sometimes after a hard run or a race.

Maybe try reducing the dose of running (duration x intensity), until you don't suffer so much, and then slowly build up if you can over weeks. You can walk & run, couch-to-5K style, if it helps.

This little-by-little approach applies to most things in running - from training your VO2max to coming back from an injury, so it may help you. The key is to understand that the timeframe of adaptations can be weeks to months, and not to rush things.

Slightly off-topic, but what exactly is this website? A "self-funded progressive news outlet... since June 2021". "Our writers are liberals, social democrats, and democratic socialists. It is not a political denomination that brings us together, rather what unites us is a desire to progress as a united country."

Sounds awfully like a front for political organisation of some stripe or other rather than a genuine news outlet.

https://ourgeneration.news/about

edit

Also kind of interesting: the "ourgeneration.news" domain was registered 11 April 2022.

https://who.is/whois/ourgeneration.news

The linked Twitter account seems to have first tweeted on 20 April 2022, doesn't get a huge amount of engagement yet already has over 15K followers:

https://twitter.com/Our__Generation

Although the article mentions "polarization", when I look at the intensity distribution graph, is that not more an example of "pyramidal" - ie with a substantial amount of work between easy and hardest (VO2) intensities (see the bars for aerobic, tempo, threshold), rather than just being divvied up between Easy/VO2 with no-mans land in the middle. (see for instance, the definitions in this paper: https://www.frontiersin.org/articles/10.3389/fphys.2015.0029...)

At the risk of whataboutism, I can’t help be reminded of that time when the British security services eavesdropped on Belgacom.

https://en.m.wikipedia.org/wiki/Operation_Socialist

There's "We should celebrate walking in Britain now more than ever" [1] from the Daily Telegraph if that's more your bag:

"At 9pm, with only the faintest glow left in the west, the birds stopped singing. Venus blazed and, as darkness deepened, others joined: Arcturus, Procyon, the Plough. A breeze riffled ink-black trees. Bats swooped. Something scuttered in the hedge. I knew exactly where I was; I’d walked this circuit of country lanes many times, and since lockdown almost daily. But I’d never done it at night. This time I wasn’t just walking with my legs; I was walking with every pore, every nerve. ..."

Could almost be written by Boris himself.

[1] https://www.telegraph.co.uk/travel/destinations/europe/unite...

Also interesting to note that Arrow itself uses FlatBuffers for its schema data.

https://arrow.apache.org/docs/format/Columnar.html

Not sure ... but didn't GSK partner with Sanofi on a vaccine that didn't work out? Sorry, I can't read your FT link :)

The AstraZeneca US phase 3 trial took about 7 months.

https://clinicaltrials.gov/ct2/show/NCT04516746

In this case, US regulators took issue with the company's press release, rather than the results of the trial itself.

Merck was the Oxford team's first choice of manufacturer. But the UK health minister overruled and told them they needed to use a British manufacturer. [1]

"[T]he Oxford-AstraZeneca vaccine was very nearly the Oxford-Merck vaccine - and under the terms of the agreement with the American pharmaceutical giant, there were no guarantees of supply.

The episode played out against the backdrop of the first phase of the pandemic. During March and April 2020, the University of Oxford negotiated a deal which would allow Merck to manufacture and distribute the vaccine it was in the process of developing.

The arrangement made sense. Unlike British-Swedish AstraZeneca, Merck had experience in making vaccines. Its senior executives had links to Oxford scientist and government adviser Sir John Bell.

Yet when the contract reached Matt Hancock's desk, the former adviser said, the health secretary refused to approve it, because it didn't include provisions specifically committing to supply the UK first."

[1] https://news.sky.com/story/covid-19-rejected-contracts-and-a...

Observational studies are a lower standard of evidence than randomised controlled trials. Countries with higher levels of vaccine skepticism (eg the US) seem to be erring on the side of doing things by the book. The UK, by comparison, enjoys a high degree of trust in its healthcare system and can afford to be a little more gung-ho.

Undoubtedly so. What UK is doing wasn't a problem until it was no longer a member.

> European politicians have been attacking the AZ vaccine because it originates from the UK and they are ideologically blinded by hatred of Brexit.

There was a good piece in the Irish Times this weekend if you fancy a less UK-centric view on this. The view you've expressed above seems to be the only one I read among the British press/commentariat. Perhaps it is the UK that is blinded?

https://www.irishtimes.com/opinion/brexit-blinds-britain-to-...

"The Oxford/AstraZeneca team managed an extraordinary feat – producing a safe and effective vaccine against a novel coronavirus within a few months. Yet time and again, AstraZeneca has undermined its own efforts with communication blunders, a lack of transparency and a consistent record of overestimating its own capacity to deliver."