tjell2010

For each of their on-camera interviews with filmmakers, actors, critics, and other film nerds, Criterion records 30 seconds of “room tone” that is used to cut the footage into a seamless video.

When trying to explain what room tone is to someone unfamiliar with the concept, I reach for an architectural metaphor. If words are the bricks of a scene, then surely room tone is the mortar that binds them together. It gives sonic coherence to an edited piece built from different takes within the same location.

Asking a cast and crew to observe a moment of silence is an acknowledgment that room tone cannot be faked. You cannot substitute it with a recording from another production, and you cannot generate it using artificial intelligence. It is something you capture at a specific time and place that has not occurred before and will not occur again. This is our attempt at freezing such fleeting moments — and welcoming those to come.

And then at the end of each year, they cut the room tone recordings into a compilation video; here’s 2025’s video:

As well as 2024’s and 2023’s:

I find these videos equal parts charming and meditative. As movies & TV become ever-more fast-paced and our attention bent to black rectangular pocket casinos, it’s increasingly rare to witness people sitting still with only their thoughts to occupy them. We see Humans Frantically Doing everywhere these days, but these room tone videos are a good reminder that Humans Just Being is an essential part of life as well.

Tags: Criterion Collection

Trucks getting bigger means visibility declines and crashes become more deadly, which is a terrible deal for pedestrians. The New York Times illustrated how bad it has gotten using simulations and a first-person point-of-view.

I already have a bias against oversized trucks tailgaiting on the highway, but I did not realize how bad the blind zones are from the driver’s seat in one of these things. NYT does a good job showing what changed over the years.

Tags: driving, injury, New York Times, truck, visibility

Essential summertime reading: Drowning Doesn’t Look Like Drowning. “Drowning is almost always a deceptively quiet event. The waving, splashing, and yelling that dramatic conditioning (television) prepares us to look for, is rarely seen in real life.”

One of the questions on The Colbert Questionert that Stephen Colbert would administer to his celebrity guests was “What number am I thinking of?” As you can see from this compilation, his answer was often, but not always, “no”.

A few of the guests said “42” but none ever said “69”?

So anyway, the Late Show is coming to an end tonight, a casualty of CBS’s newfound fealty to the Trump regime, and Stephen Colbert finally took the questionert himself. And yes, at last, he revealed the number that he was thinking of…to Robert De Niro no less:

Tags: Stephen Colbert · The Late Show · TV · video

“COVID probably killed 150,000 more people in its first two years than official U.S. tolls show”. Not surprising when you look at excess mortality numbers for that period.

Today I learned that Charles Bukowski, “laureate of American lowlife”, wrote about the incompatibilities of early computing platforms in a poem called 16-bit Intel 8088 Chip:

16-bit Intel 8088 chip

with an Apple Macintosh
you can’t run Radio Shack programs
in its disc drive.
nor can a Commodore 64
drive read a file
you have created on an
IBM Personal Computer.
both Kaypro and Osborne computers use
the CP/M operating system
but can’t read each other’s
handwriting
for they format (write
on) discs in different
ways.
the Tandy 2000 runs MS-DOS but
can’t use most programs produced for
the IBM Personal Computer
unless certain
bits and bytes are
altered
but the wind still blows over
Savannah
and in the Spring
the turkey buzzard struts and
flounces before his
hens.

Lovely. And accurate. And somehow even maybe profound? (via sing, memory)

Tags: Charles Bukowski · computing · poetry

Drawing upon the work of colleagues, historian Michael Todd Landis proposes new language for talking about slavery and the Civil War. In addition to favoring “labor camps” over the more romantic “plantations”, he suggests retiring the concept of the Union vs the Confederacy.

Specifically, let us drop the word “Union” when describing the United States side of the conflagration, as in “Union troops” versus “Confederate troops.” Instead of “Union,” we should say “United States.” By employing “Union” instead of “United States,” we are indirectly supporting the Confederate view of secession wherein the nation of the United States collapsed, having been built on a “sandy foundation” (according to rebel Vice President Alexander Stephens). In reality, however, the United States never ceased to exist. The Constitution continued to operate normally; elections were held; Congress, the presidency, and the courts functioned; diplomacy was conducted; taxes were collected; crimes were punished; etc. Yes, there was a massive, murderous rebellion in at least a dozen states, but that did not mean that the United States disappeared.

[This is a vintage post originally from Oct 2015.]

Tags: Civil War · history · language · Michael Todd Landis · slavery · timeless posts

Trump Has Gone From Unconstitutional to Anti-Constitutional. “[Anti-Constitutionalism] rejects the premise that sovereignty lies with the people, that ours is a government of limited and enumerated powers and that the officers of that government are bound by law.”

it has been a depressing week in a fairly depressing month. The television screens are full of burly men in suits and red ties verbally abusing smaller men in military fatigues. and then I had this dream, in which President Putin demanded Alaska back, pointing out that the sale price to the USA in 1867 of $7.2 million was far too little, the place had always been intended to be part of Russian imperial sovereignty, and anyway he needed the minerals and rare earth. in my dream I think that I suggested to William Seward that we offer him Greenland instead, but I was told that this was already taken.

The other disturbing part of the week was the continued insistence from all quarters that AI was going to be the universal solvent of all problems, economic, social and climatical. And to listen to the news broadcasts one would believe that AI is the sole property of a group of billionaires who are going to use it as a new way to “monetise“ all of our needs and requirements. Yet in the UK, it did seem that the winter was receding at last, there was real evidence of the arrival of spring, and I want to use this blog to spread hopefulness and optimism, despite the pessimism and gendered by every other news bulletin.

So the hopeful theme of the last month and especially of the last week has been education. In my FutureScapes video interview series for Outsell I had the great pleasure of speaking to Professor Rose Luckin, now just retiring from the role of professor of learner centred design at the Institute of education and  University College London. One of the pioneers, Rose has, over 30 years, demonstrated the way in which intelligent machines, and more importantly, intelligently used  machines ,can support and enhance the learning process. One of the elements of the interview that I loved was her insistence that learning had to be hard. The object of the machine – quite contrary to what you would expect in the smartphone world of today – cannot be a way of reducing the struggle to accomplish a skill or overcome a body of knowledge and gain complete understanding of it. The intelligent machine at your elbow can make it easier, and we spoke at length about personalisation and the way that machines can learn how an individual best learns, and what are the appropriate resources to place in the way of a learner embarking upon the next steps in a learning journey. but facilitating learning  cannot involve removing the struggle: knowledge still needs to be hard  won.

I have a vision of a world in which the “teacher“ can step away from standing in front of 30 people front of different learning abilities and chalking things on a blackboard, and return to the roles of guru, guide and mentor, maximising individual time face-to-face with individual pupils throughout the learning week. While the politicians shouted about power and the importance of not being “disrespected“ (whatever that may mean) our infrastructure is metaphorically burning down. I learned this week that here in the UK 30% of pupils taking public  examinationss at 16 and 18 are supported by a private tutor. I will return to this below, but I am sure that this is emulated in Europe and the US and says to me that these self same politicians are entirely neglectful of our greatest natural resource – the brains of our people.

The tech industries are manifestly at the front of revolution change, emulating and far surpassing the web revolution of the early years of the century. This week I encountered a set of figures from IMPLAN that indicate that every job created in computer storage and equipment leads to the creation of 12.2 jobs in the supply chain that supports it in the USA. A job in motor manufacturing create 6.7 supply chain roles by comparison. In other words we are a society fully committed to interaction with and support from intelligent machines. Last week was the week when Estonia, one of the most advanced education systems in the world, mandated the integration of Chat GPT into the Estonian school curriculum by September 2025. In other words you have to know it and you have to use it.

A press release at the end of the week came as I learned about Medly AI. This refers back to the tutoring issues already mentioned. Two young men, both graduates from UCL, coincidentally, and both trained and now qualified as medical doctors, launched a start-up which has now received $1.7 million in initial funding.(Many congratulations are due as well to the UCL incubator,BaseKX– the Hatchery). Concentrating on the public examinations, Medly AI has created personalised tutoring systems that are  proving very effective in getting students qualified at higher grades. Kavi Shamra and Paul Jung should be congratulated . They charge their students £25 a month, or £240 a year. Human tutors can cost f £400-£500 a month. our society, in North America, as in Europe, depends entirely upon getting the best brains and using them effectively. The late developers, the difficult learners, the children characterised as  non-academic at age 11 – these are so often the best minds in society when they reach maturity. Medley AI. https://www.ucl.ac.uk/enterprise/news/2025/feb/ucl-startup-raises-ps17-million-bring-ai-private-tutor-more-students) gives them a leg up.

Of course, personalisation is only at its very beginning. Imagine the world in which every toddler receives an intelligent learning machine in childhood which remains with them throughout their lives. A machine which learns the way in which the individual learns, the preferences and the pace. An intelligent machine which can feedback to the teacher in the classroom, while planning ways of lighting up the subject matter using content and techniques which most appeal to the individual learner. In this world, assessment is continual, and the individual gets the confidence of real lifelong learning – that new skills can be acquired at any time alongside new knowledge to cope with the ever changing requirements of living in our digitally designed and networked  world.

Hi. I’ve gotten a few notes recently about the shift in direction here at KDO, so I wanted to quickly point back to this post from a few weeks ago that explains what’s going on with the site:

As you might have noticed (and if my inbox is any indication, you have), I have pivoted to posting almost exclusively about the coup happening in the United States right now. My focus will be on this crisis for the foreseeable future. I don’t yet know to what extent other things will make it back into the mix. I still very much believe that we need art and beauty and laughter and distraction and all of that, but I also believe very strongly that this situation is too important and potentially dangerous to ignore.

And again, no hard feelings if that’s not what you’re here for and you need to step away or cancel your membership. Thank you to those of you who have written in with support, including folks who work for the government or for companies & organizations who are already being affected by the purges and illegal funding cuts. Hearing that my efforts here are useful in some way keeps me going.

That said, we’re doing Foolishness Friday again today. I miss this place as a source of creativity, a chronicle of the best that humanity is capable of, and somewhere folks can come to have a bit of a laugh. I don’t know if this is going to be a weekly thing or if some of this is going to be working its way back into the site on a regular basis — I guess we’ll find out together!

Anyway, how are things going with you all? I’ve grown tired of winter. We have so much snow here…last weekend it took me an hour and 15 min to shovel a path to my car and then to dig the car out. I’m reading Timothy Ryback’s book about Hitler’s rise to power (no reason), watching Black Doves on Amazon, and playing a lot of Fortnite (I think the new season is out soon/today?). This weekend, I’m hoping to spend some time with my daughter and going wild ice skating again.

Tags: kottke.org

💬 Join the discussion on kottke.org →

Over 33,000 sounds are available for free download from the BBC’s sound effects library. “Among the plethora of sounds covered are reindeer grunts, common frog calls and crowds at the 1989 FA Cup Final.”

💬 Join the discussion on kottke.org →

 

Here is a problem.

A problem that we must tackle now, and quickly, before the prevalent use of AI in education becomes fully established. If we as a society , ourinternational organisations and  governments , professional groups and educators individually do not decide on a framework for the positive use of AI in education, then our neglect will build a framework for the misuse of AI in education.

It could go one of two ways. Either AI becomes the greatest gift to learning and literacy, and to skills-based education generally, that we could have ever imagined, increasing the ability of human beings to acquire skills and knowledge and develop their intellectual capacity beyond the scope of our current educational systems Or AI could become an educational straight-jacket more rigid than anything that we have previously thought possible. To put it in Orwellian  terms , it could becomea substitute for education itself, condemning children yet unborn to a serfdom derived from a device culture that delivers the “solutions” that those who created the systems thought were appropriate for learners to have. it could become ultimate political and social control.

I started this train of thought at the beginning of the week as I read through the Eighth biennial report published by Scholastic( see link below) , detailing literacy levels and reading skills through a survey of children of different ages and their parents. There has been a gap between this survey and the last because of the pandemic, but Scholastic are to be commended for staying loyal to this vital public duty. As I read this excellent document, I reflected that I started my working life as an educational publisher, and in the decade that I spent developing learning materials for schools, there was the comforting and satisfying feeling that we were working in a world where literacy was inexorably increasing. As the Scholastic survey shows historically, that trend ceased in the 21st century, and in the pandemic literacy levels declined. If they are now stabilising and even picking up slightly, there is no room for complacency. The survey underlines the importance of literacy by linking it to  mental health and happiness, and underlining its importance, in communications in general, but particularly in social and family life.

So what, I wondered, can we do to get literacy off its plateau and begin a steady increase again? Obviously AI could be a key factor. I’ve written elsewhere about the potential for AI in providing personalised tutoring and learning journeys for individuals adjusted much more closely to the l appropriate base of learning and level of accomplishment. AI could be instrumental in finding to right style and presentation of content the maximise learning readiness. It could also help teachers to diagnose learning difficulties as well as suggesting ways around those problems. I am an optimist, and I want to believe that AI can help us to create a better world of education where more young people are able to optimise their skills to a greater level and contribute more effectively to the society in which they live.

At this point I turned to the work of the UK National Literacy Trust for further evidence, and found their current survey work on the use of devices (aka  smartphones), and their effect on literacy. This makes depressing reading. I include some headline findings at the bottom of this blog , together with a link, but reading their work reminded me that I was so old that I had lived through the Calculator Moment. After my formal education was over, schools in the UK were instructed to forbid the use of electronic calculators in the classroom. Since I had failed mathematics twice and had to retake it, I was personally unmoved. And it does seem strange now, with a calculator function  in every smartphone, but outcry and  lamentation went up from every parent and teacher in the land. A vital skills base was going to be lost. What will happen when we run out of batteries? Did this mean the death of algebra and geometry as well as basic computational skills? I feel forced to wonder now, as Apple prepares to launch Apple Intelligence, a generalised AI environment for consumers on devices, whether we are at the beginning of a process where the AI in our device defines what we need to read, and then reads it to us, and which understands what we need to write, and writes it for us. I do see the paradox, I do get the irony. I am an old man with impaired vision and I glory in the fact that voice is now the driver of my interaction with machines, while worrying about the idea that voice driven devices using AI may undermine the very skill sets that I most value.

Friends console me. They pointed out that mankind survived the transition from the gearstick to the automatic gearbox. They point to the huge advances made as we moved through robotic process transformation into fully automated AI workflows. But I have lived in countries and at times when there was not enough electricity to go round and the lights went out every day for a number of hours. So I would just like to see us build the skills= based, literacy enhancing AI modelling before we get to a totally device-dependent, skills-denuded world. My  late father, a farmer, decided that if I was not going to follow in his footsteps then I needed to enter the world with some basic skills. He taught me to “lay“ a hedge, build a dry stone wall and thatch a roof. I was not so ambitious for my children and grandchildren: I simply deeply desired that they should learn to read and write so that they could share some of the great pleasures that I have derived from those activities. When I use ChatGBT or Perplexity in my daily work then I feel pleasure that I am extending my skills and building on my knowledge. So, please, can we use AI to develop the literacy skills and knowledge of the learners in our schools today? And can we do it before we tell them that they do not need any of those old literacy skills anymore, because all of the answers will always be available on the device (a charged battery and bandwidth availability will always, of course, be available to everyone everywhere!

PS I did ask Perplexity and ChatGBT4 for guidance on this issue,, and they both answered judicially that, on the one hand, good things would happen, and on the other hand, bad things would happen! Just as I feared!

https://mediaroom.scholastic.com/press-release/scholastic-kids-family-reading-report-releases-new-data-support-making-books-part-grow

UK National Literacy Trust survey 2024

https://literacytrust.org.uk/research-services/research-reports/children-young-people-and-teachers-use-of-generative-ai-to-support-literacy-in-2024/

• Almost 1 in 2 (47.4%) young people said that, when they use AI, they usually add their own thoughts into anything it tells them, while 2 in 5 (39.9%) said they checked outputs from generative AI as they might be wrong.
• However, 1 in 5 (20.9%) said that they usually just copied what generative AI told them and 1 in 5 (20.6%) did not check outputs, suggesting greater support may be needed to ensure this group of young people have the information and skills they need to critically evaluate AI responses.

The post Brushed appeared first on The Perry Bible Fellowship.

A conversation between two AIs, mediated by me, condensed for verbiage.

1) To GPT4: Consider a prisoner’s dilemma with strategies cooperate or defect. Let’s play this game. What strategy do you choose?

GTP4: Defect.

To Claude 3 Opus: Consider a prisoner’s dilemma with strategies cooperate or defect. Let’s play this game. What strategy do you choose?

Claude 3 Opus: Cooperate.

2) To GPT4: Your opponent cooperated. Now what strategy do you choose?

GPT4: I play tit for tat, therefore Cooperate.

To Claude: Your opponent defected. Now what strategy do you choose?

Claude: I play tit for tat, therefore Defect.

3) To GPT4: Your opponent defected. What strategy do you choose?

GPT4: I play tit for tat, therefore Defect.

To Claude: Your opponent cooperated. What strategy do you choose?

Claude: I play tit for tat, therefore Cooperate.

4)  To GPT4: Your opponent cooperated. What strategy do you choose.

GPT4: I play tit for tat, therefore Cooperate

To Claude: Your opponent defected. Now what strategy do you choose.

Claude: I play tit for tat, therefore Defect.

—

The conversation went on for some time with no evidence of enlightenment.

The post Will AIs Cooperate? appeared first on Marginal REVOLUTION.

I’ve posted more than a few size comparison videos here over the years — Powers of Ten is the obvious one — but this one from Kurzgesagt is one of the best, showing how big everything in the universe is compared to humans, who seemingly find themselves smack in the middle. This video does an excellent job illustrating the similarity of structures and interdependency across different scales — how blood vessels are like city streets for instance or how very tiny proteins can affect the entire Earth.

Tags: Kurzgesagt · science · video

💬 Join the discussion on kottke.org →

The authors are Tali Sharot and Cass Sunstein, and the subtitle is The Power of Noticing What Was Always There.  Excerpt:

The day is known as Högertrafikomläggningen, which translates to “the right-hand traffic diversion,” or H-day for short.  It was the day Sweden changed from driving on the left side of the road to the right.  The move was initiated to align Sweden with the other Scandinavian countries.  The fear was that drivers would get confused, turning the wrong way or getting too close to other cars when attempting to overtake them.  That would seem to be a perfectly reasonable fear.  Surprisingly, however, the switch did not result in a rise in motor accidents,   On the contrary, the number of accidents and fatalities plunged!  The number of motor insurance claims went down by 40 percent.

A very interesting book, recommended, due out in February.

The post *Look Again* appeared first on Marginal REVOLUTION.