RSS Feeds

После очередного запуска компьютера привычные программы одна за другой перестали открываться. В папке C:\custom остались в основном пустые каталоги: исчезли проекты, программы и другие данные общим объёмом около 800 ГБ.

Я начал запускать приложения по одному и смотреть, после какого действия пропадают файлы. След привёл к Telegram Desktop — мессенджер, установленный в другой папке, при каждом запуске рекурсивно очищал C:\custom.

Дальше были Process Monitor, стеки вызовов, issue на GitHub и неожиданная причина в проверке орфографии.

Хороший маркетинговый анализ — основа практически любого нового проекта. Пока не понимаешь, кто твоя аудитория, кто конкуренты, какие у них предложения и цены, двигаться дальше просто некуда. Раньше на такой анализ уходило недели две-три, а иногда и целый месяц ручной работы. Сейчас этот процесс можно ускорить в разы благодаря нейросетям. 

Привет! Меня зовут Дмитрий Сатаров, я руководитель агентства «Эй, Стартапер!» и  наставник курса «Нейросети для маркетинга» в Яндекс Практикуме. Когда в 2022–2023 годах появились неплохие нейросети, я начал выстраивать процесс, который позволял бы выполнять маркетинговые исследования значительно быстрее. Поделюсь им в этой статье.

Не откладывай на завтра то, что можно автоматизировать сегодня. Сотрудникам IT-компаний, которые работают по ТК или сервисному договору, зарплата приходит без их участия: платежами занимается бухгалтерия. Но если хочется уйти на фриланс, с финансами придётся иметь дело самостоятельно.

В способах приёма денег от клиентов режим налога на профессиональный доход (НПД) не ограничивает: подойдёт оплата через СБП, по номеру карты, QR-коду и платёжной ссылке. Но есть нюансы: чеки, налоги, лимиты, возвраты. Именно поэтому в статье рассказываю, как принимать оплату самозанятым в IT и как автоматизировать рутину и снизить риск ошибок при формировании счетов и чеков.

Проверка на переполнение есть в коде, проходит ревью и работает в отладочной сборке — а после оптимизации исчезает из бинарника. Разберём, почему компилятор имеет на это полное право, как неопределённое поведение влияет на указатели, память и проверки, и чем ловить такие ошибки до того, как они проявятся в рабочей среде.

Каждый день мы натыкаемся на одни и те же грабли в связке фронта и бэка. Каждые такие грабли — это деньги и время. А если ещё и договариваться не умеем, процессов нет и решать никто не берётся, то дальше начинаются конфликты, атмосфера в команде проседает и проблемы копятся ещё быстрее.

Статья не про саму автогенерацию — с ней всё понятно, про неё и без меня написано достаточно. Мы внедрили генерацию типов из OpenAPI, и она вытащила наружу то, что с генерацией напрямую не связано: кто не хочет писать комментарии, почему у одной сущности два имени и почему у нас падал dev. По сути — про команду, эго и умение договариваться.

Сразу оговорюсь: это не единственно верный путь. Команды и компании разные, у всех свои нюансы, я рассказываю, как было у нас. Но если автогенерации у вас ещё нет и вы соберётесь её внедрять — будете хотя бы знать, что вас ждёт.

Unpopular new policy comes as midterms loom and control of Congress could be decided by several Michigan races

Every day $1bn worth of goods crosses the river dividing Detroit, Michigan, and Windsor, Ontario, two largely blue-collar US and Canadian cities that have come to act as one in the creation of North America’s auto industry.

News this week of Donald Trump’s escalating trade spat with Canada has everyone worried. Political and economic leaders and observers labeled the move an act of “hubris” , blamed Trump’s “ego”, and called the escalation “insanity”.

Привет, Хабр! Будем знакомы — grizzzer, участник различных CTF-команд и независимый исследователь кибербезопасности. В этой статье разберу, как я реализовал критическое событие в банковском сегменте онлайн-полигона Standoff 365 — «Дефейс веб-сервиса ДБО First Partner Bank».

Покажу всю цепочку — от DNS-разведки до эксплуатации уязвимости server-side template injection (SSTI) в шаблонизаторе Pug и финальной подмены текста на главной странице.

The UK prime minister clearly still adores the band. Perhaps he’s attracted to songs about sham justice and murderous meat factories, shot through with nihilism and exquisite despair

As an origin story, Andy Burnham: Smiths Fan has a lot going for it. It’s plainly authentic: he can name the time he saw them live (Salford University, Queen Is Dead tour, 1986). A cynical young man in a hurry might fake that now, but not back when he first said it, in 2013. Asked by Labour Uncut for his desert island tracks during his 2010 leadership bid, three of the eight were sung by Morrissey (two Smiths singles and one off Viva Hate), which is superfan territory.

That’s a softball exercise to endear yourself to Labour members; you have eight chances to hit every base, and to blow three lives on an artist who had already, by then, been dabbling in nationalism for 18 years (we’ll come back to this) was dicey and brave. You’ll recall that Burnham did not win that election. If you were an enemy state, amassing this hinterland for an asset, you’d have basically done a really good job.

As well as being quieter, cheaper to run and easier to clean, induction cookers can cut kitchen temperatures by up to 10C

Leaping flames, searing heat, copious sweat and scorchingly inventive swearing – restaurant kitchens are never a place for the faint-hearted, but this summer they have been sweltering in successive heatwaves to a dangerous degree.

Some restaurants, cafes and bakeries in the UK grew so hot that they were forced to close, or reduce their hours, for the safety of their staff. There is still no maximum working temperature, despite calls from the TUC, and tens of thousands of kitchen staff are thought to have carried on working in “furnace-like” conditions.

Может ли статистика помочь маркетологам и отделам продаж в обработке лидов? Да, причем в целом ряде задач!

Прочитав статью, вы:

1. Получите мощный аргумент для переговоров: «Нет, вероятность, что конверсия вашего отдела продаж просела случайно – статистически ничтожна»;

2. Оцените вероятности целевых событий для вашего медиаплана;

3. Прокачаете «План/Факт», построив эталонное распределение.

As well as being quieter, cheaper to run and easier to clean, induction cookers can cut kitchen temperatures by up to 10C

Leaping flames, searing heat, copious sweat and scorchingly inventive swearing – restaurant kitchens are never a place for the faint-hearted, but this summer they have been sweltering in successive heatwaves to a dangerous degree.

Some restaurants, cafes and bakeries in the UK grew so hot that they were forced to close, or reduce their hours, for the safety of their staff. There is still no maximum working temperature, despite calls from the TUC, and tens of thousands of kitchen staff are thought to have carried on working in “furnace-like” conditions.

Unpopular new policy comes as midterms loom and control of Congress could be decided by several Michigan races

Every day $1bn worth of goods crosses the river dividing Detroit, Michigan, and Windsor, Ontario, two largely blue-collar US and Canadian cities that have come to act as one in the creation of North America’s auto industry.

News this week of Donald Trump’s escalating trade spat with Canada has everyone worried. Political and economic leaders and observers labeled the move an act of “hubris” , blamed Trump’s “ego”, and called the escalation “insanity”.

Привет, друзья! Одним из двигателей моей любви к моделям, к исследованиям и ко всему, чем я сейчас занимаюсь, является математика. Математические структуры удивительны — открыли ли мы математику или создали — слишком сложный вопрос. Я верю, что открыли и мне очень нравится наблюдать, как математика возникает в моделях.

Так появилось это исследование и так появилась карта. Задача, которую я поставила на входе "Оценить насколько воспроизводима и эквивалента в LM-моделях математика". Что получилось — смотрим в статье.

С этого материала хочу начать небольшую серию квартальных обзоров рынка конструкторов сайтов и CMS. Сервисы сейчас развиваются настолько быстро, что уследить за всеми релизами, сменой подходов и новыми технологиями становится все сложнее. Поэтому раз в несколько месяцев буду собирать наиболее заметные изменения в одном месте и смотреть не только на то, какие функции появились, но и на то, куда в целом движутся сами платформы. Этот летний дайджест будет первым выпуском такой серии.

===

Лето 2026 года оказалось довольно насыщенным для рынка конструкторов сайтов. За три месяца накопилось достаточно изменений, чтобы посмотреть уже не на отдельные функции, а на то, куда движется рынок в целом.

Если посмотреть что было сделано практически у любого мало-мальски крупного сервиса, там по-прежнему можно увидеть много привычных вещей: новые блоки, шаблоны, интеграции, SEO-настройки, локализация, формы и тд. Однако наиболее интересные релизы последних трех месяцев происходили в несколько другой плоскости.

uCoz дополнил MCP средой для запуска собственного серверного кода. Яндекс начал принимать заказы от продавцов, которым вообще не нужен собственный интернет-магазин. Framer пустил AI-агента непосредственно в canvas. Webflow начал решать вопрос не только о том, что агенту разрешить, но и как изолировать его изменения от production. WordPress продолжил собирать инфраструктуру, через которую возможностями CMS смогут пользоваться внешние AI-клиенты.

Общий знаменатель этих релизов – AI постепенно перестает быть только генератором и получает доступ к действиям внутри самой платформы.

A heartbroken woman becomes fixated on a famous writer in this witty, touching follow-up to Sorrow and Bliss

The protagonist of Meg Mason’s new book, the follow-up to the 2020 tragicomedy Sorrow and Bliss, is an invisible woman. She might well be standing there but you won’t notice her. She’s a techie at literary festivals – the person who shuffles up to a famous writer a few minutes before their event and apologetically slides a microphone cable down their shirt while they’re making small talk with the chair or accepting a bottle of room-temperature mineral water from an anxious publicist.

No man is a hero to his valet, and no writer is a hero to the technical staff lurking in the wings of a venue. Mason’s opening pages present a teasing but not quite libellous series of cameos from real writers: George Saunders deciding at an inconsiderately late stage to insist on swapping his headset for a lapel mic; Ann Patchett being frosty with a fan; Mick Herron accidentally stepping on someone’s foot and not apologising.

Poetry with wonder and wit; a heartwarming tale for horse lovers; a portal to the future; a brilliant YA wartime romance and more

Small Monster, Big Monster by Allan Sanders, Phaidon, £7.95
Refrigerator is EMPTY. Monster is FULL (of pizza!). An eye-wateringly bright board book, featuring splendidly zany, toddler-pleasing pairs of opposites.

The Flying Machine of Dreams by Kate Wakeling, illustrated by Viviane Schwarz, Otter-Barry, £12.99
A delectable picture-book collaboration that perfectly distils the wonder of the everyday, from the textures of a leaf to a squishy-cushioned den. Wakeling’s irresistible poetry is richly heightened by Schwarz’s sweet, strange illustrations.

Nepalis and Tibetans must not pay the price – with their lives, livelihoods and money meant for education and health – for global heating caused by wealthy states

• Sunil Acharya is an international climate policy and just transition specialist based in Kathmandu, Nepal

As a Nepali citizen working on international climate policy for more than a decade, it is difficult to look at the harrowing images coming out of the catastrophic flash flood that thundered down valleys in Nepal, sweeping down the Lhende Khola, Bhotekoshi and Trishuli rivers, and not feel angry.

Not just because of the destruction, but because it comes after years of warnings from scientists and civil society that a rapidly warming planet would make disasters more frequent, more intense and harder to predict. This is what the cost of delaying global climate action looks like.

Sunil Acharya is an international climate policy and just transition specialist based in Kathmandu, Nepal

Do you have an opinion on the issues raised in this article? If you would like to submit a response of up to 300 words by email to be considered for publication in our letters section, please click here.

Строчка, которая встречается в скриптах повсеместно:

mysql -u admin -pSecretPass123 -e "select 1"

Пока команда выполняется, этот пароль видит каждый, у кого есть доступ к устройству — обычный, непривилегированный. Достаточно посмотреть список процессов.

Привет! Я Костя Никитин, руководитель направления в одном из подразделений Т-Банка. Занимаюсь проектированием высоконагруженных систем — мы импортозамещаем автоматизированную банковскую систему. Если упростить, это автоматизация бухгалтерского учета: все банковские операции приходят к нам, нужно их сложить и посчитать. В Java-мире я уже больше двенадцати лет, начинал когда-то с 1.6.

Расскажу о проблемах, с которыми мы столкнулись, когда переехали в Kubernetes, как мы их порешали и как в итоге оптимизировали ресурсы при работе в виртуализации. Если вы хоть раз ловили OOMKilled на ровном месте и не понимали за что, может быть, узнаете себя.

Спойлер: в финале мы сократили потребление памяти на 18%, CPU — на 31%, а приложения при этом стали работать быстрее. Погнали.

Men in the tech industry are covering up their legs, even as their counterparts in finance are increasingly open to showing them off.

(Image credit: SIBAS_minich/iStockphoto)

As the death toll rises from flash floods that hit the Nepal-Tibet border, experts are warning more flooding could be on the way. And, Fed Chairman Kevin Warsh is expected to give a key address today.

(Image credit: Prakash Mathema)

Health insurance enrollment through the Affordable Care Act is plummeting.

Nearly 3 million fewer Americans now have insurance through the program, popularly known as Obamacare, compared to last year, according to federal data released in June. After congressional Republicans allowed enhanced subsidies for the program to expire in January, many Americans saw their health insurance premiums spike by an average of about 30 percent. 

As a result, many Americans have been forced to drop their coverage entirely.

With affordability among the top concerns on voters’ minds heading into the midterms, healthcare costs have become a central issue for candidates.

“A lot of policy experts say, look, when you raise the price of something, demand goes down. And that’s what’s happening here,” KFF Health News senior correspondent Julie Appleby told Today, Explained. “It’s the prices — people are dropping out.”

Appleby has been reporting on healthcare policy for decades, and she joined Today, Explained co-host Noel King to break down the latest changes to the Affordable Care Act, why enrollment is plummeting, and how the ACA crisis is teeing up healthcare as a central issue in the upcoming midterms. 

Below is an excerpt of their conversation, edited for length and clarity. There’s much more in the full podcast, so listen to Today, Explained wherever you get podcasts, including Apple Podcasts, Pandora, and Spotify.

Sixteen years ago, President Obama signed the Affordable Care Act into law. Julie, what was the situation at the time that made it feel so necessary? What did the Affordable Care Act aim to do? 

There was a lot going on. We had a lot of people who were uninsured. We had premiums going up. Many times in the past there had been some effort to do some kind of health reform, but nothing ever really passed. We had Medicare and Medicaid in the ’60s, but no major legislation since then.

People had employer coverage, like they do now, but there were a lot of people who were on their own and buying in what’s called the individual market.

The individual market was really tough. They could tell people, “Hey, you’ve got a health condition, we’re not taking you.” They could you could sign up for an individual plan and later, if you did get some kind of health condition, they could go back through your medical records and see, “Hey, did they tell us about every single thing that had happened to them in the past five or ten years?” And if not, they could retroactively cancel your plan.

What did the Affordable Care Act do exactly?

It allowed young people to stay on their parents’ insurance plan up to age 26, and that was a really big deal and also a very popular provision in the Affordable Care Act.

It created these subsidies to help people purchase coverage on the individual market. It pays for part of your premium. After you pay a certain percentage of your household income, the subsidies kick in and it picks up the rest, up to a certain up to a certain income level. 

It did some other things, like they can’t tell you, “You have a preexisting condition, so we’re not going to take you.” They’ve got to take you.

It created a benefit package that covers all these essential things like hospitalization and drug costs and maternity care, which might not have been in plans before. They used to be able to charge women more than men. They can’t do that anymore.

What was the conversation around passing the Affordable Care Act? Remind us how and why it got so toxic and pitched.

It was super polarized, kind of like what we have now. There was a lot of discussion around whether this is going to be too much government or if it was just the right amount. A lot of people thought, “No, this is terrible, it’s gonna destroy things.”

And I think polls reflected that for the first four or five years or more of the Affordable Care Act. It was pretty evenly divided, when you looked at opinion polls, between people who thought it was a good thing and people who didn’t think it was a good thing.

In more recent years, that has changed and the favorability ratings have gone up for the Affordable Care Act and now majorities do favor the Affordable Care Act.

If you had told me that 16 years ago, I would not have believed you. Here we are. And how is the Affordable Care Act doing?

To use medical terminology, is it on life support? No, it’s actually doing fairly well. But there are some symptoms that there are problems. 

Remember at the end of last year, there was all the discussion about these things called enhanced premium tax credits and whether they were going to be extended or not. The government shut down over this. Those were not extended. As a result, it costs more this year for most people in the plan, and what we’ve seen is during open enrollment last year we had about a million people fewer sign up and that was about a five percent drop. 

That was 5 percent just in signups at the end of last year. Since then, the first few months of the year, we’ve seen about 3 million people drop off. Some people are finding that they can’t afford it and they are not paying their bills and they are getting dropped. A 3-million-person drop-off is pretty substantial.

There’s also a group of folks, particularly conservatives and Republicans, who say, “You know what, the numbers were inflated.” There was all this improper enrollment. People were getting these zero-premium plans and so there was fraud, there was improper enrollment, the subsidies were too generous or there were other factors that led to some fraudulent activity. And so these drop-offs just represent people who were fraudulently enrolled dropping out for one reason or another. 

There’s this real split right now between what are the reasons for the drop off. But most policy folks, most people who study the situation here say the main reason is people just aren’t able to afford it.

Why is it so expensive? Why are healthcare costs so high? 

That’s the $64,000 question, right? 

And if you get it right, you win. You’re in charge.

That’s been the question since I’ve been covering this debate. Look, it’s not just Affordable Care Act costs going up. Employer plans are going up. Everything’s going up. 

Generally, healthcare costs go up faster than inflation and that’s driven by a lot of things. It’s driven by drug costs, it’s driven by hospital costs, it’s driven by labor costs. So all of those things go into the mix with the Affordable Care Act. But also this year, there’s this additional factor of the subsidies that went away. 

Insurers were looking at that saying, “You know what, the subsidies are going down. We expect to have lower enrollments next year. The people most likely to stay enrolled are the ones who have health problems because they don’t want to take the risk of being uninsured.” So insurers were calculating that in when they set their premium rates for this year: We might have sicker people in our pool that we’re gonna have to cover. 

That’s part of what we’ve seen with costs in the Affordable Care Act. But everybody is seeing health inflation go up right now and that’s affecting premiums.

Here we are 16 years after the passage of the Affordable Care Act. Has it accomplished what it set out to do?

You have to look at that in a number of ways. It aimed to expand coverage and reduce the uninsured. I think the answer there is yes, the number of uninsured has gone from 13 or 14 percent to as low as just under 8 percent a couple of years ago. It’s ticked up a little bit since then.

Did it control costs? That’s a harder one to answer. And it’s back to that $64,000 question. How do you do that? Costs have continued to go up. Premiums are definitely higher than they were pre-ACA. To be fair, the plans are very different. They cover more, they’re more generous, they’re better coverage, but they are more expensive.

The big question is: What is going to be put forward as an alternative and how palatable will that be to consumers, lawmakers, the healthcare industry? It’s going to be challenging, but I think we’re going to hear a lot more discussion this year. Whether or not we see actual proposals that could pass, I think that’s a little more questionable.

У банков и ритейлеров, с которыми я работала, почти всегда одна и та же боль: данные о клиенте размазаны по CRM, core-banking и платёжным системам, и никто не может ответить на простой вопрос — а что мы вообще знаем об этом клиенте прямо сейчас. Чтобы попрактиковаться в решении этой задачи и показать подход публично, я собрала pet-проект: end-to-end пайплайн, который строит единую витрину Customer 360 из синтетических данных.

Сразу оговорюсь: все данные генерируются локально скриптом и полностью синтетические. Кода или данных работодателя в проекте нет — это самостоятельная реализация архитектуры, которую я использую в повседневной работе DWH/data-аналитика.