❌

Reading view

Tech sell-off drives US markets down by nearly $1tn after subpar earnings reports

Big tech stocks falter as investors question whether billions being poured into AI can generate returns quickly enough

A sell-off of tech stocks that began earlier this month and gained steam with disappointing earnings from leading tech companies drove US markets down by hundreds of billions of dollars Friday. In total, the decline has knocked nearly $1tn in value off the tech-heavy Nasdaq exchange in one week and erased about $2tn in a month.

The S&P 500, the index of the 500 most valuable companies in the US, fell by more than 1% Friday, losing hundreds of billions of dollars in market value. It was on track to erase its gain for October, even though it set its all-time high during the same month. The weakness for technology stocks dragged the Nasdaq composite down even further, more than 3%, also erasing hundreds of billions in value. The Down Jones Industrial Average fell less.

Continue reading...

© Photograph: Brendan McDermid/Reuters

© Photograph: Brendan McDermid/Reuters

© Photograph: Brendan McDermid/Reuters

  •  

Tech sell-off drives US markets down by nearly $1tn after subpar earnings reports

Big tech stocks falter as investors question whether billions being poured into AI can generate returns quickly enough

A sell-off of tech stocks that began earlier this month and gained steam with disappointing earnings from leading tech companies drove US markets down by hundreds of billions of dollars Friday. In total, the decline has knocked nearly $1tn in value off the tech-heavy Nasdaq exchange in one week and erased about $2tn in a month.

The S&P 500, the index of the 500 most valuable companies in the US, fell by more than 1% Friday, losing hundreds of billions of dollars in market value. It was on track to erase its gain for October, even though it set its all-time high during the same month. The weakness for technology stocks dragged the Nasdaq composite down even further, more than 3%, also erasing hundreds of billions in value. The Down Jones Industrial Average fell less.

Continue reading...

© Photograph: Brendan McDermid/Reuters

© Photograph: Brendan McDermid/Reuters

© Photograph: Brendan McDermid/Reuters

  •  

Amazon gives Wall Street mixed financial picture in earnings report

E-commerce company posts high cloud computing sales but gives a cautious sales forecast for the fourth quarter

Amazon offered a mixed picture of its financial health on Thursday in its quarterly earnings report.

The e-commerce company gave a cautious sales forecast for the current fourth quarter, which encompasses the critical holiday shopping period, after delivering fiscal third-quarter results that came in below Wall Street analysts’ expectations.

Continue reading...

© Photograph: Charles-McClintock Wilson/NurPhoto/Shutterstock

© Photograph: Charles-McClintock Wilson/NurPhoto/Shutterstock

© Photograph: Charles-McClintock Wilson/NurPhoto/Shutterstock

  •  

«Искусственный» Луки Гуаданьино: сюжет, сроки и проблемы проката

Значительную роль в развитии современных подходов к машинному обучению сыграл французский исследователь Ян Лекун. В 1989 году он и его коллеги из AT&T Bell Labs представили свёрточную нейросеть, которая распознавала рукописные цифры. Систему обучали на тысячах изображений цифр из настоящих почтовых отправлений в США. В том же году разработчики описали результаты в научной статье (doi:10.1162/neco.1989.1.4.541).

Более того, эти разработки внедрили для реального использования. В девяностых на основе этих лабораторных экспериментов создали коммерческие системы распознавания банковских чеков. В статье 1998 года исследователи отчитывались об обработке нескольких миллионов чеков ежедневно (doi:10.1109/5.726791). Как рассказывал Лекун, к концу девяностых такие системы считывали от 10 до 20 % всех чеков в США.

Достижения отмечали как наградами в отрасли, так и славой вне неё: Лекун получил премию Тьюринга за 2018 год, а в «Книге рекордов Гиннесса» за Bell Labs зафиксирован рекорд по созданию первой нейросети, способной распознавать рукописные символы. В декабре 2024 года учёный не без удовольствия выложил на личном канале YouTube восстановленную видеодемонстрацию нейросети LeNet 1 от 1989 года.

Читать далее
  •  
❌