scapbi

After sharing my story, I realized how widespread this has become. Almost every friend I know has at least one relative who has been a victim. These are not one or two isolated cases. It is almost everyone. This is no longer an exception. It is becoming normal, and that is terrifying.

One urgent warning I want to share is about how scammers are now using AI.

In my country, they are already using AI for video calls, voice cloning, and highly scripted conversations. The person on the screen can look and sound real. The conversation feels natural and authoritative. For many people, especially older relatives, it is almost impossible to tell that it is fake.

This is no longer about obvious scam messages or bad phone calls. AI is making scams feel personal and trustworthy. Please warn your families. No bank or authority will ever ask for OTP codes or urgent transfers, even if it is a video call and the person looks real.

Thank you, @tomhow. The quick response and the welcoming attitude from the moderators are something I personally appreciate a lot.

As I mentioned before, HN is a technology community. I hope that one day I can share only technical topics here, so we can focus the discussion on technical matters. I hope we can stand together as engineers, doing our part and taking responsibility, even if that does not fully address the root causes.

Thanks everyone for your kind words and advice. It really means a lot to me right now. I hope someday I can get over this and turn it into a lesson, and I will share what I learned so others can be more careful.

To @dang and the moderators: I honestly do not understand why this topic was flagged. This situation is real and very painful for my family, and it is beyond my imagination that sharing it would be a problem.

Your last story is deeply moving. It captures the human cost behind these crimes in a way that statistics never can. Her quiet strength, expressed in just a few words, is a powerful reminder that even after devastating loss, people can find the will to endure and move forward.

My main background is in mathematics, and after 15 years working in big data, I am now focusing on machine learning and AI. Not flashy technologies, but practical ones that can do things beyond what a normal human can do alone.

As for the human side, I have already shared my story within a closed community where other engineers know me personally. I am not begging for help; my goal is to inform more people about this situation and, if I am fortunate, to find others who are willing to join me in this fight.

It is almost impossible. Unlike in developed countries, where banks can offer some level of protection to customers, in third-world countries, banks mainly protect themselves. All responsibility is pushed onto users. Banks take no accountability, and the government protects the banks.

Let me give a concrete example. When money is transferred to scammer accounts, it is immediately distributed across hundreds of other accounts and moved out of the electronic system in under 30 seconds. At that point, everything is gone.

It is much easier than you might imagine. Hundreds of thousands of people live together almost like a city, scamming their own people, speaking the same language as their victims. They train day and night to manipulate the human mind, supported by an underlying system of fake accounts. When I was building the system against that, I felt like David facing Goliath. The problem goes far beyond what technical systems alone can solve.

Thank you very much for your kind words. What I am thinking about most right now is, first, encouraging my mother and helping her through this difficult situation. Beyond my responsibility to my family, I also feel a responsibility related to the system I built. I want to connect with people who have much more knowledge than I do and see whether we can do something meaningful for third-world countries. That has always been my lifelong dream.

I know about corruption; anyone who lives in a third-world country understands it. However, as an engineer who built the technical solution, I feel hopeless. Even though my team has strong technical skills and worked extremely hard, with very little sleep for six months, to create something for the greater good, situations like this still happen. In fact, they are becoming worse and more frequent.

As an engineer, I hope that I can gain more knowledge, connect with more people, and do more to help those who have no one to protect them.

Thanks for your kind words.

Thanks for your kind words! Most math textbooks are packed with formulas and abstract stuff, so I tried to do something different, make it fun while still getting the ideas across. Hope you like it!

Crafting Interpreters https://craftinginterpreters.com is a super friendly, step-by-step guide to building your own language and VM, looking forward to seeing what kind of language you come up with too!

We’ve been building Mochi, a small statically typed language for querying structured data like JSON, CSV, and in-memory lists. In the latest release (v0.10.5), you can write SQL/LINQ-style queries with from, where, join, and group by, all fully type-checked and testable inline.

What makes it different: queries compile to a register-based bytecode VM (written in Go) with compiler passes like constant folding, liveness analysis, and dead code elimination. It’s meant to feel like writing SQL inside a real language, backed by actual compiler infrastructure.

If you’re interested in how query engines, DSLs, or virtual machines work, this might be worth exploring.

This thread was the final push I needed to add logic programming to Mochi https://github.com/mochilang/mochi — a small statically typed scripting language I’m building for agents and real-time data.

I gave OpenAI Codex a single prompt with a sample like:

  fact parent("Alice", "Bob")
  rule grandparent(x, z) :- parent(x, y), parent(y, z)
  let gps = query grandparent(x, z)

And it generated a working Datalog engine in Go with:

  - fact storage + recursive rule support
  - bottom-up fixpoint evaluation
  - unification and `!=` constraints
  - FFI bindings to expose `fact`, `rule`, and `query` to scripts

Full thinking process: https://chatgpt.com/s/cd_684d3e3c59c08191b20c49ad97b66e01

Total implementation was ~250 LOC. Genuinely amazed how effective the LLM was at helping bootstrap a real logic layer in one go.

The PR is here https://github.com/mochilang/mochi/pull/616