🚩2023 MetaTrust Web3 Security CTF

🔗 ctf.metatrust.io

🖋Registration will end in 3 hours Sep 14 10:00:00pm GMT+8

📄 Competition will end in 15 hours Sep 15 10:00:00am GMT+8

#CTF

TinyML and Efficient Deep Learning Computing
6.5940 • Fall 2023 • MIT

🔗 efficientml.ai

Large generative models (e.g., large language models, diffusion models) have shown remarkable performance, but they require a massive amount of computational resources. To make them more accessible, it is crucial to improve their efficiency.

This course will introduce efficient AI computing techniques that enable powerful deep learning applications on resource-constrained devices. Topics include model compression, pruning, quantization, neural architecture search, distributed training, data/model parallelism, gradient compression, and on-device fine-tuning. It also introduces application-specific acceleration techniques for large language models, diffusion models, video recognition, and point cloud. This course will also cover topics about quantum machine learning. Students will get hands-on experience deploying large language models (e.g., LLaMA 2) on a laptop.

🖼 The slides and lab assignments from the last semester are available for access here.

🙋 Speakers and mentors
🌕Instructor: Song Han, Associate Professor, MIT EECS
🌕TA: Ji Lin, PhD Student, MIT EECS
🌕TA: Han Cai, PhD Student, MIT EECS

💻 Format: online — lecture recordings are available at YouTube.
⏲ Frequency: Live Streaming — Lectures are live streamed at live.efficientml.ai every Tuesday/Thursday 3:35-5:00pm Eastern Time.
💬 Discussion: Discord
🎓 Homework submission: Canvas
💳 Resources: MIT HAN Lab, HAN Lab Github, TinyML, MCUNet, OFA, SmoothQuant

TON Blockchain course 🎓

This course will be dedicated to developing on The Open Network (TON) blockchain.

The TON blockchain was launched by the founders of Telegram, Nikolai and Pavel Durov, and later transferred to the open TON community that has been supporting and improving it since then. The Open Network allows processing of millions of transactions per second. The TON ecosystem includes blockchain, payment platform, TON Proxy, file storage, and other decentralized services.

The course was designed to be as close to reality as possible so that participants can work on their own projects on the TON blockchain, or get a job at existing startups and funds.

This course includes
🌕46 lessons
🌕9 hours 6minutes of video
🌕55 quizzes

🔗 stepik.org/course/176754/promo#toc

MixBytes Farm открывает набор на курс «Аудитор смарт-контрактов»

Подойдёт всем, кто хочет работать с низкоуровневым кодом и эффективными алгоритмами, разбираться в логике DeFi или освоить новую профессию.

🗓 Старт обучения: 30.10.23  
🎓 Прием заявок и проверка тестов: до 27.10.23 (включительно)
🥇 Стоимость: бесплатно
👒 Условие поступления: входное тестирование
⏲ Длительность: 1.5 месяца 

Программа:

🌕Ethereum architecture;
🌕Solidity low-level patterns;
🌕Solidity high-level patterns;
🌕DeFi protocols code review;
🌕DeFi typical vulnerabilities;
🌕DeFi hacks;
🌕CTF (capture the flag); 
🌕Test audit.

*лекции размещены в записи, поэтому ты сможешь проходить теорию в удобное время дня. 

После успешного завершения обучения выдается сертификат в виде SBT (soulbound token). Он не только позволит участвовать в закрытых аудит-контестах по поиску уязвимостей в составе команды Stronghold, которая образовалась из выпускников MixBytes Farm, но и станет опорой при поиске работы аудитором смарт-контрактов.

🖋 Зарегистрироваться и пройти входной тест 👈

🙋 Для связи @kudryaviyartemiy

🔗 tondevstudy.io

Your mission: Improve the open-source bot solution for liquidations from the EVAA team, using existing protocols.

🗓 Important dates:
Development phase: from November 27 to December 14
Robot competition: from December 15 to December 21
Evaluation and awarding of winners: from December 22 to December 30

🎁 There are valuable prizes

🎓 All information is in hackathon documentation of the hackathon

📷 How Liquidations Work on DeFi? B.Protocol explained part 1

@evaabuidl

💬 @megdu_skobok
🔗 youtube.com/watch?v=wwPmcDihxFs
Игорь Корсаков и Сергей Тихомиров обсудили множество аспектов блокчейна и его будущего. Вот краткое и организованное резюме их обсуждения:

🌕Определение и Понимание Блокчейна
- Блокчейн как Технология: Описывается как распределенная база данных с цифровыми подписями, характеризующаяся неэффективностью и плохой масштабируемостью.
- Важность Блокчейна: Подчеркивается его роль в распределении власти и прав на запись данных.

🌕Историческая Значимость Биткоина
- Биткоин как Исторический Документ: Обсуждается важность биткоина для понимания архитектуры блокчейна.
- Сборка Технологий в Биткоине: Биткоин собрал вместе существующие технологии для создания децентрализованных денег.

🌕Технологические Аспекты Блокчейна
- Элементы Блокчейна: Рассматриваются децентрализованный консенсус и безопасность.
- Алгоритмы Консенсуса: Обсуждаются Proof of Work и Proof of Stake.

🌕Применение Блокчейна
- Ограниченное Применение: Указывается на отсутствие успешных примеров применения блокчейна вне сферы криптовалют.
- Блокчейн и Цифровые Подписи: Пример с Rolex показывает, что иногда достаточно цифровых подписей и централизованной базы данных.

🌕Будущее Блокчейна
- Пессимистический Взгляд: Выражается мнение, что блокчейны могут быть не так важны для общества, как это представляется энтузиастам.
- Разделение Экосистемы: Предполагается возможное разделение на "трушных" энтузиастов блокчейна и более компромиссные системы.

🌕Советы и Рекомендации
- Изучение Компьютерных Наук: Рекомендуется изучать компьютерные науки, так как блокчейн включает в себя множество базовых компьютерных задач.
- Внимательность к Новым Проектам: Советуется быть внимательным к новым проектам в области блокчейна и пытаться понять их суть.

В целом, обсуждение охватывает широкий спектр тем, связанных с блокчейном, от его основных принципов до будущего развития и практического применения.

🎤 @w3voice?videochat
🗓 сегодня, начнем примерно после 17:00 UTC или когда соберутся участники
💬 тема TON Smart Challenge #5

🔗 arxiv.org/abs/2310.20216

This research by Cameron Jones and Benjamin Bergen from UC San Diego evaluates GPT-4's ability to pass the Turing Test in a public online experiment. The study found that the best-performing version of GPT-4 successfully passed the test in 41% of cases, outperforming previous models such as ELIZA (27%) and GPT-3.5 (14%), but did not reach the random level and the level set by human participants (63%).

The Turing Test, devised in 1950, is a game where one human and one artificial intelligence try to convince an interviewer that they are human using only a text interface. This test was developed as an indirect way to answer the question, "Can machines think?".

In the experiment, 652 people participated, playing a total of 1810 games. Of these games, 1405 were analyzed by the researchers (260 involving a human as a witness). Participants were randomly assigned to play either with another human or with artificial intelligence. AI games were conducted in the absence of an available human partner.

Analysis of the results found that participants' demographic data, including age, gender, education level, frequency of interaction with chatbots, and level of knowledge about large language models (LLMs), did not affect the accuracy of their judgments regarding whether their interlocutor was human or AI.

Among the strategies used by interviewers to determine whether the interlocutor was AI or human were small talk, questions about knowledge and reasoning, and situational awareness. The main reasons interviewers decided their interlocutor was AI were related to linguistic style and socio-emotional factors.

The study raises the question of whether 30% is an adequate criterion for passing the Turing Test, noting that a more widely accepted criterion is 50%. It is noted that none of the tested AI witnesses achieved either 50% success or parity with human witnesses. Thus, the study does not find evidence that GPT-4 passes the Turing Test.

#ML

📊 How many developers are there in the world?

🌕On GitHub, by January 2023, there were more than 100 million active developers registered [1].
🌕According to the latest data, in 2023, there were more than 500,000 active developers registered on the WakaTime platform, who spent a total of 51 million hours programming with WakaTime plugins. On average, each user spent 48 minutes per day programming, and less than 10,000 users contributed 50,330 hours of Solidity use. [2][3].
🌕There are 226,399 developers participating in the cryptocurrency and Web3 ecosystems every month, with 6,279 active full-time as of October 2023 [4].

Based on this data, we can make a rough guess that at least 0.23% of developers are working on Web3

#report

The Unison programming language represents a rather unique and innovative tool in the world of software development:

1. Immutable Codebase: In Unison, code is stored as abstract syntax trees (ASTs), not text files. This allows for easier version management and avoids many problems associated with traditional version control systems.

2. Simplified Refactoring: Since Unison identifies functions by their content, not name, refactoring becomes simpler and safer. Changing a function automatically propagates to all places of its use.

3. Distributed and Parallel Programming: Unison was initially developed with support for distributed and parallel code execution in mind, making it suitable for cloud computing and high-performance computing systems.

4. Type Safety: Unison's strict type system provides additional reliability and helps prevent common errors during compilation.

5. Simplified Testing: In Unison, tests can be directly linked to functions, simplifying code maintenance and quality assurance.

Areas of application for Unison include, but are not limited to, the following:

- Cloud Application Development: With support for distributed code execution, Unison is well-suited for creating scalable cloud applications.

- Systems Requiring High Reliability: Strict typing and simplified refactoring make Unison an attractive choice for systems where code reliability and safety are important.

- Scientific and Research Projects: For projects requiring complex calculations and distributed data processing, Unison can be useful due to its parallel programming capabilities.

- Education: Unison's unique approach to code management and typing can be useful for educational purposes to teach programming concepts and version control systems.

Thus, Unison offers a range of interesting possibilities that can be useful in various fields of programming, especially where reliability, scalability, and ease of code management are important.

🔗 unison-lang.org
🔗 exercism.org/tracks/unison/exercises

#language

🗓 18:00 CET, 7 Декабря — будет встреча на тему TVM-Solidity
🎤 @EverQuorum

🔗 chain.link/developers/experts

#Chainlink #program

🔗 encode.club/lightlink-hackathon

#lightlink #hackathon

🔗 dorahacks.io/hackathon/abstractathon/detail

#XION #hackathon

🗓 сегодня 25 января в 16:00 CET 🎤 @EverQuorum

Гость Illya Gerasymchuk https://illya.sh/

🌕ZK: история, текущий статус, перспективы
🌕Как и какие блокчейны L1, L2 используют ZK
🌕Проект zkLocus: Private Geolocation Off & On-Chain - история, текущий статус, планы
🌕На встрече будем планировать воркшоп на протоколе Mina и где-то еще где это возможно, который состоится на следующей встрече