Pull to refresh

My feed

Type
Rating limit
Level of difficulty
Warning
To set up filters sign in or sign up
Article

What's New in the Angie 1.9 Web Server (an nginx fork) and What to Expect from 1.10?

Level of difficultyEasy
Reading time8 min
Views139

You may have already read in the news that on the eve of Cosmonautics Day, a new stable release of Angie 1.9.0 was released, an nginx fork that continues to be developed by the team of former nginx developers. Approximately every quarter, we try to release new stable versions and delight users with numerous improvements. This release is no exception, but it's one thing to read a dry changelog and quite another to get to know the functionality in more detail, to learn how and in which cases it can be applied.

The list of innovations that we will discuss in more detail:

— Saving shared memory zones with cache index to disk;
— Persistent switching to a backup group of proxied servers;
— 0-RTT in the stream module;
— New busy status for proxied servers in the built-in statistics API;
— Improvements to the ACME module, which allows automatic obtaining of Let's Encrypt TLS certificates and others;
— Caching TLS certificates when using variables.

Read more
Article

Captcha Solver Extension – Which to Choose: AI-Powered or Human-Powered? Difference Free and Paid CAPTCHA Extension

Level of difficultyEasy
Reading time14 min
Views76

How Does a Developer Realize They Need a Browser Auto CAPTCHA Extension?

Imagine a developer automating routine tasks — for example, testing a web application or writing a data scraping script. Everything runs smoothly until a CAPTCHA appears on the path. In the browser, a familiar window pops up: "I am not a robot," or a grid of images where you need to find traffic lights or pedestrian crossings. The automatic script halts, tests fail, and an inexperienced developer might not even realize the problem for a long time — after all, they set everything up and started it, but didn’t account for the presence of CAPTCHAs (Completely Automated Public Turing test to tell Computers and Humans Apart), designed precisely to stop bots. But what if the bot is ours and performs, say, useful work?

Read more
Article

How to catch and optimize problematic queries in PostgreSQL

Level of difficultyMedium
Reading time8 min
Views201

If you work with PostgreSQL, you've likely run into performance issues at some point — especially as your database grows. Things may have been running smoothly at first, but as your client database expanded, queries started slowing down. Sound familiar? Here's a guide to help you identify and fix problematic queries, so you can get your PostgreSQL database running at peak performance again.

Read more
Post

У меня есть знакомый энтузиаст LLM, который также изучает верилог. Я попросил его написать инструкцию к упражнению с неким сенсором, который он интегрировал. Он разумеется сбросил это на LLM, я почитал и понял, что LLM нужно запретить как распостранение Экстази и "солей" среди молодежи. Точно так же как "дизайнерские наркотики" дают ощущение счастья и достижения без труда, сгенеренная LLM документация выглядит как реальная, вот только читателю она не поможет.

Что нужно читателю? Картинку как подцепить сенсор к плате, временную диаграмму сигналов которые от него выходят и пару слов про проблемы, которые у него возникнут (дребезг) и как их стоит решать. Так чтобы было достаточно информации, чтобы сесть и написать код на верилоге.

Что выдал LLM? Сначала пять абзацев мутного словестного описания что "изменения переключателей проходят некоторую последовательность, позволяющую определить направление", с галлюцинациями что движется и что неподвижно. Потом не имеющую отношения к задаче информацию, из каких материалов делаются эти сенсоры в разных странах мира, чтобы быть дешевыми для хоббистов и образовательных учреждений. Далее про разные способы решения проблемы дребезга, в том числе способы, не имеющие отношения к данной ситуации. И наконец, куски определения пинов из QSF и XDC файлов из случайных примеров в интернете, которые не имеют отношения к описанному примеру, так как в нем во-первых эти файлы не используются (другой вендор, другой способ задания пинов), а во-вторых, в нем эта часть проекта абстрагирована (пользователю вообше это не нужно это делать).

То есть текст просто водит читателя за нос, не давая ему никакой полезной информации для решения проблемы. Но даже это не важно, потому что читатель этот текст читать не будет, так как учует LLM в заголовке и убедится в третьем предложении, после чего перестанет читать. Текст является иллюстрацией терминов "сделать на отцепись" и "из дерьма и палок".

UPD: И самое страшное: это 27 страниц вместо 1 страницы полезной инструкции, которую я ожидал. ДВАДЦАТЬ СЕМЬ СТРАНИЦ ЛАБУДЫ !!!

Я хочу обратно в годы, когда этого ужаса не было. Нашу цивилизацию ждут тяжелые времена. Я уже видел в ЖЖ посты агитирующие на перевод всей порноиндустрии на generative AI.

Tags:
+16
Comments18
Article

FunCaptcha (Arkose Labs) solver: Principles of Operation, Features, and Methods for Automated Bypass

Level of difficultyEasy
Reading time12 min
Views800

We continue our journey through the world of CAPTCHAs (Fantastic CAPTCHAs and Where to Find Them, as well as Methods to Combat Them), and today we encounter yet another “tough nut” in the CAPTCHA universe – FunCaptcha (Arkose Labs).

FunCaptcha is a type of CAPTCHA developed by Arkose Labs that offers users small puzzles instead of the usual tasks like recognizing distorted text or selecting images containing buses. In traditional CAPTCHAs (e.g., reCAPTCHA), verification often relies on recognizing distorted characters or simple images. Arkose Labs took a different route: their “entertaining” CAPTCHAs feature interactive challenges with 3D objects, logic puzzles, and audio questions. This approach is intended to be user-friendly for humans while complicating life for bots.

Typical FunCaptcha challenges include:

Read more
Article

Tips and methods for conducting user research with children

Level of difficultyEasy
Reading time5 min
Views918

When creating an app or any other software product for children, one of the hardest parts of the process is conducting user research correctly. Although working with kids might seem fun and entertaining, it takes certain skills to get them engaged in testing your product and voicing their opinion so you could gather all the necessary information. It’s important to understand that mentally kids function differently than adults, so working with a young target audience requires a different approach. Treating children like adults in the UX research process can lead to serious mistakes: they might not get a proper understanding of your product and you might end up getting wrong results, only wasting your time and budget. In order to avoid that, we’ve collected a few tips below that might help you communicate with kids more effectively for a productive and fruitful research session.

Read more
Article

One-click Postgres Pro optimization with pgpro_tune

Level of difficultyEasy
Reading time4 min
Views477

Imagine a familiar situation: it’s Monday morning, tasks are piling up, and you need to quickly spin up a new service using Postgres Pro. Or maybe you’ve just upgraded your database server over the weekend — added more CPUs, more RAM.

Here’s how to get your database tuned and ready to make the most of the new hardware and workload, without wasting time.

Read more
Article

Database performance analysis using pg_profile and pgpro_pwr

Level of difficultyEasy
Reading time4 min
Views132

DBAs often struggle to identify the most resource-hungry processes that degrade system performance. Back in 2017, DBA — and now Postgres Professional engineer — Andrey Zubkov faced the same challenge. This led him to develop pg_profile for PostgreSQL, which has since evolved into pgpro_pwr.

In this article, we’ll dive into strategic database monitoring and show you how to pinpoint bottlenecks in your databases using our tools.

Read more
Post

ChatGPT: LOOKING FOR A CAFFEINE SUBSTITUTE

Although caffeine stimulates mental activity and aids in eSports/games and late-night programming, it has many side effects, such as increased blood pressure and crazy heart rate, a sharp rise and quick drop in stimulation. Therefore, we need a high-quality alternative to caffeine, and we will search for it using AI. Potential candidates to replace caffeine are Theacrine (or TeaCrine) and N-Phenethyldimethylamine Citrate (USA FDA said Ok).

ChatGPT successfully created a very complex table, even with a calculated column based on FUZZY criteria (if you can do this in SQL — you're a genius!), but it struggled with sorting the table. Attention: there is an image below, links are not clickable.

this is an image, links are not clickable
this is an image, links are not clickable

In the first numeric column, it failed to sort the numbers in descending order. I spent about 15-20 minutes trying. I experimented with various prompts and explanations. This is strange.

This tool (ChatGPT) understands table manipulation commands very well. In this example, I asked it to create a table based on data from large stores, specified which columns were needed and what information they should contain, indicated the order of the columns, including relative positioning — for instance, "insert a column with such-and-such data before this column" — and even more.

IT was able to create a SUMMARY column based on previously generated columns — this is the column with weighted sums of substance weights from other columns, and IT independently found the weighting coefficients quite accurately.

Moreover, for each product, IT managed to identify the substance composition based on specific criteria and listed them, creating a separate column. Not all substances, but only those filtered by certain criteria (only those that are not caffeine but have an effect similar to caffeine — try programming such a query in SQL manually without AI, taking into account the fuzzy criterion of similarity of effects, and also determine the similarity coefficient for creating the weighted sum of substance masses per serving of the dietary supplement). And it even partially managed to sort by the weighted sum.

But despite completing so much complex work, it still made a small mistake with sorting.

Tags:
0
Comments0
Article

How I Created Perfect Wiki and Reached $250K in Annual Revenue Without Investors

Level of difficultyEasy
Reading time6 min
Views86K

Hi, my name is Ilia. I founded Perfect Wiki — a SaaS product for creating internal company knowledge bases that works directly within Microsoft Teams. We created a simple and convenient tool for storing, editing, and sharing knowledge within companies. It all started with the idea to resolve one specific pain point: the built-in Wiki in Microsoft Teams offered was inconvenient, and there was no worthy alternatives with full integration to the platform.

In this article, I want to share how the idea came about, the mistakes I made, how I found my first customers, and how I gradually grew to a steady income of $250,000 a year over five years. All of this — without investors, a 20-person team, or a “Series A” round.

Read more
Article

How to Fail Those Students Who Rely on ChatGPT

Reading time3 min
Views2K

We at Verilog Meetup constructed an exam/interview problem that has an interesting property: if a student tries to figure out a solution by thinking by himself, he usually succeeds; however if he dumps the problem on ChatGPT, the solution fails (does not pass the automated test), and the student goes into a death spiral of futility, kicking ChatGPT to get the solution right.

There is nothing weird about the problem, we do this in the industry all the time:

Read more
Article

Google ADK: Easiest Way to Build an AI Agent

Level of difficultyEasy
Reading time7 min
Views1.9K

In this tutorial, I’ll explain in simple terms what AI, AI agents, and workflows are, and then I’ll walk you through building your very first AI agent in Python using Google’s Agent Development Kit (ADK). By the end, you’ll understand the differences between these concepts and have a working content-assistant agent you can run from your terminal or a web interface.

Read more
Article

By next year, we'll be talking to databases in natural language

Level of difficultyEasy
Reading time4 min
Views633

According to Gartner, natural language queries will replace SQL as early as 2026. 

While Gartner's prediction may be optimistic, the shift toward natural language interfaces for databases is inevitable. The timeline may vary, but the transition itself is a certainty.

Read more
Article

What is DPI Engine?

Level of difficultyMedium
Reading time25 min
Views1.3K

For people familiar with the term DPI (Deep Packet Inspection), it often carries an unpleasant association: blocking, regulators, censorship, tightening controls. In reality, DPI is simply the name of a technology whose essence lies in the deep analysis of network traffic. Deep traffic analysis involves identifying protocols, extracting the most significant fields and metadata, classifying internet services, and analyzing the nature of network flows. I will explain how such solutions work in this article.

Read more
1
23 ...