RSS Feeds

Image captured in early June shows animal foraging in the waterway in first since early 1900s, signaling its recovery

An otter has been spotted in the Bronx River for the first time in more than 100 years, in a further sign that New York’s previously befouled waterways are recovering some of their ecological health.

The North American river otter was photographed by a motion-activated camera foraging near some vegetation in the Bronx River, a waterway that flows through New York’s Bronx and empties into the East River.

Discovery marks latest grim finding in lake as water levels in crucial reservoir outside Las Vegas continue to decline

Human remains have once again been discovered at Lake Mead outside Las Vegas, Nevada, marking the latest in a line of gruesome discoveries as water levels in the crucial reservoir continue to drop, authorities said.

These remains were discovered on 16 August, USA Today reported. The National Park Service (NPS) told the outlet there did not appear to be anything suspicious about the body.

Military orders criminal investigation into deaths of five-year-old and six family members in Gaza in 2024

The Israeli military admitted on Wednesday its troops had opened fire on a car carrying five-year-old Hind Rajab and six family members in Gaza in 2024, saying it had ordered a criminal investigation into the killing.

The killing of Hind and her family was one of the most high profile of a series of incidents involving civilian casualties during the first months of the Gaza conflict which prompted international outrage.

Сравнений WordPress и Tilda в интернете уже достаточно. Обычно они сводятся примерно к одному и тому же набору критериев: стоимость, простота освоения, скорость создания сайта, SEO, количество шаблонов и необходимость заниматься техническим обслуживанием.

В результате вывод тоже часто получается предсказуемым: Tilda проще и быстрее, а WordPress сложнее, зато гибче.

Но если посмотреть на эти платформы глазами веб-разработчика, разница находится немного в другом месте.

Она становится особенно заметной, когда сайт перестаёт быть набором из пяти-десяти страниц и на нём появляются повторяющиеся данные: услуги, проекты, сотрудники, объекты, товары, характеристики, отзывы. А затем эти данные начинают зависеть друг от друга.

Поэтому в этой статье я хочу сравнить WordPress и Tilda не столько по внешним возможностям, сколько по их внутренней логике:

Military orders criminal investigation into deaths of five-year-old and six family members in Gaza in 2024

The Israeli military admitted on Wednesday its troops had opened fire on a car carrying five-year-old Hind Rajab and six family members in Gaza in 2024, saying it had ordered a criminal investigation into the killing.

The killing of Hind and her family was one of the most high profile of a series of incidents involving civilian casualties during the first months of the Gaza conflict which prompted international outrage.

Я разработчик, и за годы работы несколько раз пытался подобрать систему планирования, которая не превращалась бы в ещё один проект по обслуживанию самой себя.

Первая попытка была ещё в университете. Я купил большую тетрадь, расчерчивал её по дням и записывал, что хочу сделать. Закончил дело — отметил. Не закончил — перенёс дальше.

Какое-то время система работала.

Было приятно открыть неделю и увидеть, чем я вообще занимался. День не растворялся где-то между учёбой, интернетом и бытовыми делами.

Но довольно быстро появилась другая проблема.

Каждое утро я снова писал примерно одно и то же. Зарядка, английский, тренировка, какие-то обычные дела. Потом часть не успевал сделать, переносил на следующий день и постепенно начинал чувствовать, что уже обслуживаю собственный ежедневник.

В какой-то момент я его просто забросил.

Через много лет попробовал снова. Тогда у меня был тяжёлый период на работе, и хотелось хотя бы следить за тем, чтобы вне работы оставалась жизнь.

Я записывал прогулки, отдых, фильмы, английский, звонки близким.

На обложке блокнота ещё нарисовал календарь и отмечал каждый день смайликом. Что-то вроде 🙂, 😐 или 🙁.

В конце месяца можно было одним взглядом посмотреть, как он прошёл.

Мне это нравилось.

Но и этот блокнот через несколько месяцев отправился на полку.

Go 1.27 (на момент написания rc3, финал обещают в августе) приносит generic-методы: метод наконец может объявить собственный тип-параметр. Взял golang:1.27rc3 в Docker и проверил, где у фичи границы. reflect новый метод не видит вообще, хотя go/types на том же типе показывает его в method set — метод доступен компилятору и недоступен через рефлексию, поэтому ни один интерфейс им не удовлетворить, а валидаторы и ORM его не заметят. Померил цену: на 200 инстанцирований с запретом инлайна бинарь с методом больше на 2264 байта, из которых 2000 приходятся на длину имён символов, а не на код; по времени разницы нет — это generic-функция с ресивером и тот же gcshape-стенсилинг. А вот что реально заметит прод на апгрейде: GODEBUG=asynctimerchan=1 теперь роняет процесс до main, encoding/json v1 на движке v2 меняет байты вывода на битом UTF-8, json.RawMessage обзавёлся String() и поменял ваши логи. Плюс замеры нового аллокатора: минус четверть времени на new(16 B) и ноль на 128 байтах.

Presidential deputy chief of staff Iryna Mudra dismissed day after former defence minister called for wartime election

Ukraine’s president Volodymyr Zelenskyy has fired a senior aide amid a fresh corruption investigation – one day after a former defence minister warned the country was facing “a crisis of governance” and needed to hold a wartime presidential election.

Iryna Mudra, a presidential deputy chief of staff, was dismissed by decree after multiple reports in Ukrainian media that her house was being searched by officials investigating an alleged money-laundering scheme.

Presidential deputy chief of staff Iryna Mudra dismissed day after former defence minister called for wartime election

Ukraine’s president Volodymyr Zelenskyy has fired a senior aide amid a fresh corruption investigation – one day after a former defence minister warned the country was facing “a crisis of governance” and needed to hold a wartime presidential election.

Iryna Mudra, a presidential deputy chief of staff, was dismissed by decree after multiple reports in Ukrainian media that her house was being searched by officials investigating an alleged money-laundering scheme.

Есть такая старая программистская байка, что нельзя платить программистам за строки кода, так как тогда они будут писать длинный бестолковый код и любить метод copy-paste. Будущее наступило. Только теперь этими “программистами” является генеративный ИИ (GenAI), которому как раз платят за строки кода. Иронично.

Один из моих интересов — изучение сгенерированного С++ кода, чтобы понимать, как развивается индустрия создания ПО, какие проблемы уходят, а какие наоборот возникают. После заметки “Дайте посмотреть на нормальный С++ проект, созданный вайб-кодингом” мне предложили заглянуть в проект VibeTensor, что я и сделал.

Всем привет! Меня зовут Заруцкий Семён, я выпускник и новоприбывший аспирант физического факультета МГУ. Изначально я занимался экспериментальными квантовыми технологиями, со временем появилась необходимость моделировать физические процессы на компьютере, а следом в мою жизнь вошли прикладные ML‑проекты. Пришло всё к тому, что недавно мне посчастливилось принять участие в летней школе по искусственному интеллекту «Лето с AIRI» 2026, которую Институт провёл в Первом университетском лицее им. Н.И. Лобачевского в Усть‑Лабинске совместно с Фондом «Вольное дело». 

О том, как прошла школа и что интересного успело произойти, я расскажу в этой статье. Делать это я буду не один, мне поможет Настя Золотарева, сотрудник Института ИИ МФТИ, биофизик по образованию и моя коллега по школе. Мы оба прошли через это увлекательное и полное депривации сна приключение, а теперь хотим поделиться им с вами. Если вкратце, это был интенсив на грани возможностей, но в потрясающе тёплой атмосфере. 

А теперь с самого начала.

1000 сайтов в час вместо нескольких часов на каждый — так работает ЮScan, сервис автоматизированного аудита от ЮMoney. Его основная задача — помочь банкам и платёжным организациям не допустить подключения сайтов, которые маскируют запрещённую или рискованную деятельность под легальный бизнес. Сервис автоматически проверяет ресурс и выявляет подобные случаи ещё до начала сотрудничества.

С 2020 года через ЮScan проверено более 550 тысяч заявок. За это время ЮKassa не получила штрафов за обслуживание запрещённых видов деятельности.

Кому это нужно? Если вы обрабатываете более 10 000 страниц в сутки, работаете с эквайрингом, маркетплейсами, KYC/AML-проверками или выполняете требования 152-ФЗ, ручные процессы неизбежно становятся узким местом. Меня зовут Ирина, я руковожу антифродом в ЮMoney. Расскажу, как устроен ЮScan: как обходить тысячи страниц, находить рискованный контент и превращать данные в понятный скоринг.

People told they and their pets should keep out of river at Ham amid anger over Thames Water’s treated sewage plan

Swimmers are being warned not to enter the water at the first official bathing site on the River Thames, in south-west London, because of toxic blue-green algae in the water.

The blue-green algal blooms are rare in moving water but are likely to be caused by the extreme heat, low river flow, and nutrient pollution from sewage waste.

There's a reason why the escapist hit from singer Aitana has swept most of the Spanish-language world.

(Image credit: Pablo Alzaga)

When we retrain, when we rebuild, and when we leave a model alone.

By: Harrison Katz

A forecast that carries weight

The Forecasting Data Science team at Airbnb produces many of the forecasts the rest of the company plans around: demand, bookings, cancellations, and a range of finer cuts by market and segment, refreshed continuously across thousands of markets. The targets differ, and the models differ, but they have one thing in common: Other teams build on top of them.

This means a forecast that is casually wrong is not a clean miss, as it might be in an academic setting. That’s because a small bias does not stay small once a lot of decisions are riding on it. So when one of those forecasts starts to drift, what to do about it is not really a methods question. It is a risk question, and an easy one to get wrong, which we have from time to time.

One of these forecasts had been missing, compared to what actually happened after the forecast was released, in the same direction for a couple of quarters. This bias persisted after several routine refreshes. The usual solution would be to fully retrain the model: pull in the recent data, refit the model again, and ship. But we wanted to understand the source of the bias, rather than simply hoping an update would eliminate it.

If you’re interested in other posts on this topic, you can learn more about how COVID impacted Airbnb’s financial models or how we dealt with disruption to our models during the pandemic. This post is about the discipline that came out of both: how we now decide whether a struggling forecast needs new data, a new model, or no changes at all.

One word, three decisions

The easy mistake is treating the choice to “retrain” a model as a single action. It is three separate actions — refitting, respecifying, or holding — and none of them is particularly similar to the others.

Refitting is the cheaper option. Same model, same structure, same features, updated with newer data. This is what most people mean by “retrain.” For the ordinary drift that builds up in a model as the world moves on, it’s usually the right choice.

Saying that this is the cheaper option is not the same as saying that it’s cheap. A refit still has to be validated and shipped. The refitting process can degrade a model that was actually fine if the recent data happens to be unusual. And, given the importance of the production forecast, every refresh is a small risk you are choosing to take. It is the least expensive of the three options, but not a free one. None of this is an argument against refitting. On most cycles it is the right move, and if we had to run one of the three blind, it would be this one. The claim is narrower. A refit is a decision with a price, and pricing it is what lets you notice the cycles where a different option was worth more.

Respecifying is a different animal. You change the model itself: add a feature, drop one, change the structure, the priors, the likelihood. On the production forecast, which so many of our teams plan around, this is a real commitment; you are replacing something you understand, and have watched the behavior of for years, with something partly new that you have not. It is also where almost all the actual improvement lives. Refitting keeps a good model current. Respecifying is how a model that is in some way wrong in its operations is made right.

Holding is the one nobody likes. You look at the miss, decide it does not warrant action, and leave the model alone. This takes the most nerve, because to anyone watching the forecast it looks like you are ignoring a problem. On a forecast that people are relying on, “We decided to do nothing” is a hard sentence to say out loud. It is also, more often than you would think, the correct call.

What we have learned from years of running these models is that the trouble comes from filing all three under one verb. “We should retrain” gets said when any of the three is the actual answer, and the word drags everyone toward the one that feels cheapest, because it is the one with a name.

Most of the time, though, nobody says it at all. Retraining is rarely a decision someone makes because a model is drifting. Retraining runs on a schedule, monthly or quarterly, by convention, because a standing cadence is one less thing to think about. Which means the choice among the three is hardly ever made on purpose. The cadence makes the choice in advance, every cycle, and the cadence always picks refit.

Three ways to hold on too long

Each of the options contained in the term “retrain” — refit, respecify, or hold — has a matching failure, and underneath each sits the same mistake: a model holding onto a shock, or a surprise, after the cause of the shock or surprise is no longer relevant to forecasts. That mistake can grab a blip and treat it as the new normal; keep a crisis assumption after the crisis is over; or build the shock so far into a new model that the model can never move past the shock.

Chasing noise. A forecast misses for a quarter, the residuals look alarming, and someone asks whether we should update the model now rather than wait for the next scheduled run. That impulse is the failure, and it is a specific one. What triggers an off-cycle refit is always a surprise, which means the window you are rushing to absorb is the window you understand least.

A couple of years ago one of our markets ran hot for a quarter against the published forecast, because a large event landed on the calendar and pulled a wave of bookings forward. We refit early. The model took the unusual quarter as the new level, the next two forecasts came in high, and it settled only once the event aged out of the training window. Waiting for the cadence would have meant fitting that spike alongside the quarters that came after it, rather than as the last thing the model saw, which is a much weaker pull.

Carrying ghosts. The quieter error, and the opposite mistake. A structural assumption that was true during a shock stays switched on long after the shock has passed, hidden behind a run of refits that all look like routine maintenance. We have had one of these hide in plain sight. During the worst of the COVID disruption, cancellation timing shifted in a way it never previously had, and the model learned from it. Market behavior came back to normal well before the assumption did, and since every refit looked clean, the stale piece sat there until the forecast had leaned the same way for long enough that someone went looking, found the lean baked into the model, and respecified the model to wring out the now-faulty assumption.

Respec-as-panic. The failure of overcorrecting. Something genuinely moves — a foreign exchange rate (FX) swing from a major source market, a regional disruption that reroutes demand — and the reflex is to rebuild. A fresh model, new structure, stood up under deadline.

We have caught ourselves reaching for this solution. An FX move on a major source market threw one of our forecasts off for a couple of months, and the first instinct was to stand up a new model around the new regime. We widened the priors instead, the existing model rode it out, and we avoided a move to a more fragile model, tuned in response to a shock that was already on its way out. (Figure 3)

How we decide now

The rule we now hold to is easy to state and hard to follow. On a standing cadence, the question of whether to retrain mostly answers itself. So the question that matters is which of the three options a given miss calls for: refit, respecify, or hold? And the right answers turns on what changed.

If the process generating the data is still the one the model assumes, and the parameters have just drifted, refit. The structure is fine, it just needs current numbers.

If the process has changed in a way the model cannot represent no matter how you estimate it, respecify, because fresh data cannot help a model being asked to describe a world it has no language for. The tell is direction: a misspecified model misses the same way over and over.

A respecify can cut both ways. It can add structure the model was missing, the way borrowing across geography did after 2020, or it can take structure out, retiring an assumption the world has outgrown.

The second kind is the forgetting, and it is the one teams skip, because adding feels like progress and removing feels like giving something up. But it’s an important tool to remember, and to use, whenever needed.

And if the miss sits inside the range the model already calls normal, hold. In practice this usually means declining to refit off-cycle, on the theory that a surprise you cannot yet explain is the worst possible reason to move a model early. The cadence will get to it. Skipping a scheduled refit is the rarer and more expensive version, since stale parameters have a price of their own, and it is worth it only when you can name the thing in the window that you do not want the model to absorb.

There is a more formal version of this that we have written up separately [1][2]. The one-line version is that updating is not free: moving a model toward new data costs something, staying put costs something, and the call is a comparison of the two, at the level of parameters and again at the level of structure.

In production, running real forecasts, the math is not what we reach for. We reach for the three questions above, in order.

The time we got it right

The clearest case we have of making this call correctly is the one from our second post on the Covid era, seen through this lens.

Before 2020, our destination-market forecasts leaned on a hierarchy. Markets with long, stable histories anchored the estimates, and other, “thinner” markets borrowed strength from them, on the assumption that a destination behaves like comparable destinations elsewhere. The borrowing was the structure, and for years it was true enough to be both invisible and unquestioned.

COVID broke the assumption, not the parameters. The shock was larger than anything we had previously experienced. Recovery did not arrive everywhere at once, or in the same shape. Some markets came back fast, while others stayed flat for quarters. Markets s that used to move together were suddenly on different paths. As a result, the “borrowing” of one market’s stability, by other markets that had previously tracked it, which had previously stabilized the forecasts, was now contaminating them. Markets were being pulled toward a kind of average that no longer described any of them.

The instinct was to refit our models with new data, and we tried that. It failed in a way that turned out to be the whole lesson: the estimates did not just shift, they went unstable, swinging quarter to quarter as the model tried to reconcile markets that no longer belonged in the same pool. Over the recovery window, the refit-only error ran about three times the pre-shock baseline, and it would not settle as new data came in.

That instability was the signal. A parameter problem gets noisier at the edges, but holds its shape; this was the shape itself coming apart, which is what a structural problem looks like.

The fix was a respecification, which is the subject of the previous post, so in one line: we changed what the model borrows across. Instead of pooling by a fixed hierarchy, the prior borrows along geographic adjacency and shared recovery dynamics, so a market draws strength from places actually behaving like it today, rather than places it used to resemble before the crisis. On a held-out recovery window, the respecification cut error by a little over half against refitting, and brought it back to within a couple of points of the pre-shock baseline.

The point for this post is that our new framework, which gives a prominent role to holding, would have told us not to refit. That would have prevented the failed refit that generated so many errors.

The miss was directional, not noisy, so it was not a hold. Fresh data made it worse, not better, which is the signature of a problem in structure and not in parameters, which indicated that the remedy was not a refit. What was left was to respecify.

We got there by trial, and errors, that time. The reason to write the decision down is to get there without the trial the next time, on a forecast where the trial costs a quarter of accuracy that other teams feel downstream.

Why this is hard on forecasts that carry weight

None of this is hard to understand. It is hard to do, and it is hardest precisely on the forecasts that matter most, for three reasons that we keep running into.

Holding looks like negligence. When a forecast a lot of people watch is missing, “We are choosing not to act” is an unpopular thing to say, and the pressure is always toward visible activity. Instead, when we hold, we use the modern, Buddhist-adjacent maxim: “Don’t just do something; sit there.” Refitting, on the other hand, is visible activity. It photographs well even when it is wrong.

Respecifying is expensive, and a little frightening. Replacing the structure of a model that has produced reliable numbers for years means giving up something you understand for something you do not understand yet, and the cost of being wrong is not abstract. That fear is healthy, and it is also why teams under-respecify and let ghosts accumulate.

And the last few years trained the wrong reflex. Forecasting through COVID meant changing models under real pressure, often correctly. Coming out the other side of the pandemic, a lot of teams kept reaching for the rebuild when a refit would do, while leaving pandemic-era assumptions switched on because nobody wanted to touch a model that had survived.

The same period produced both the panic and the ghosts. The discipline now is not how to react fast. It is how to tell the three options apart again, with time to think, on models where being repeatedly wrong is expensive.

Learning to forget

Refitting keeps the model current. It does not keep it honest. What goes stale is not the numbers but the assumptions underneath them, and those are exactly what a refit leaves alone.

A shock makes you add structure to cope: an elevated risk, a wider band, a special case for a world that has stopped behaving. Taking that structure back down once the shock has passed, before it quietly biases everything downstream, is harder than any retrain. It rarely has a deadline. Nobody schedules it.

A good forecasting team doesn’t just learn from shocks. It learns to forget them, when needed, on purpose.

If this type of work interests you, check out some of our related positions.

Acknowledgments

Thanks to Liz Medina, Jess Needleman, Linhan Liu, and Flavio Stanchi for building and improving the forecasting systems described here, and to Peter Coles, Yuanyuan Cui, and Adam Liss for their support of this work and its publication.

Harrison Katz leads Finance Data Science & Strategy at Airbnb. His research focuses on Bayesian methods for compositional and hierarchical time series, Bayesian decision theory, & forecast governance.

All product names, logos, and brands are property of their respective owners. All company, product, and service names used in this website are for identification purposes only. Use of these names, logos, and brands does not imply endorsement.

How we knew COVID was over (and what our models had to unlearn) was originally published in The Airbnb Tech Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.

Про ИИ-ботов пишут в основном пересказами документации вендоров: этот собирает данные для обучения, тот ходит за содержанием по запросу пользователя. Я решил посмотреть, что происходит на самом деле — прежде всего по журналам доступа сервера. Ниже представляю вам разбор: кого видно в логах, что они запрашивают, чем отличаются два типа ботов у одного вендора и почему файл llms.txt в этой картине почти не встречается.

Reverse decluttering is here, prompting a shudder from me and my fellow neat freaks. Why would you want to be surrounded by mementos when you could be perfectly tidy?

Older doesn’t always mean wiser. Jo Whiley has spoken about a home hack she picked up from – clang! – Dame Judi Dench, and I’m afraid I’m opposed to it with every fibre of my being. Whiley is in a new phase – her four children are grownup – so she recently began the most satisfying project of all: a decluttering spree. Hard even to type those words without a shiver of anticipation. Trigger warning for fellow neat freaks, though: this anecdote does not end well.

Whiley went to interview Dench for her RHS Roots podcast and found the actor’s house “packed with family photographs, children’s drawings and memories everywhere”. Mind-blowingly, she means this as a positive. “Seeing Judi’s reminded me that lived-in spaces are often the most beautiful because they reflect a lifetime of love, family and shared experiences,” she says. “I suddenly thought: ‘Why am I putting all these things away?’” As soon as I got home, I immediately put all the family photographs back out.” TLDR: I cleaned up and then I made it all messy again. Shudder.

The maker of indie classics Kids, Swoon and Poison on daring to have a temper in the 90s, what some producer credits mean – and what happened between her and Joaquin Phoenix

Trailblazing producer Christine Vachon bestrides the world of US independent film-making like a colossus. She has championed idiosyncratic American cinema for more than 35 years, scoring hits such as the romcom Materialists, which starred three superhero actors (Dakota Johnson, Pedro Pascal, Chris Evans) and grossed $108m, alongside curiosities such as Zola, a strippers’ road-trip comedy based on a string of viral tweets.

Vachon got her start in the early 1990s. Her often scandalous movies – from Poison (gay prison sex) and Swoon (gay child killers) to Happiness (paedophile dad) and Kids (delinquent teens) – dominated the cultural conversation decades before boutique studio A24 had sold its first tote bag. Her production outfit Killer Films, which she founded in 1995 with Pamela Koffler, has as its logo a cartoon bunny with a target for a body, mischievously evoking the company’s blend of irreverence and danger. “I don’t think we’ve ever made a movie that somebody hasn’t been offended by,” Vachon said in 2000.

A weak yuan guarantees Beijing massive trade surplus even as the White House tries tamp down on Chinese imports

Since China offered a truce in the trade war last October following its threat to deprive the United States of rare-earth magnets, the Trump administration has been happy to stop escalating. After all, imports from China have been falling, down by 40% in the year to June, compared with the same period in 2024. Best to declare victory and call it a day.

But the US hasn’t won this battle – and there’s strong evidence to show it’s too chicken to really try.

У каждого есть навык или черта, которые его выделяют среди других, определяют многое в его поведении и в том числе то почему он занимается тем делом, которым занимается. Есть подтип программистов, которым присуща педантичность, аккуратность, желание привести все в порядок, расставить по своим местам, в общем упорядочить хаос. Я как раз такой. Если вы тоже из “этих”, то прошу под кат, обсудим как с этим жить :)