Own the harness, own the agent orchestrators.
Great to see open-source work starting to enable it.
Being able to compose and combine multiple agents is clearly the future to avoid model lock-in. Curious how routing works, as that remains unsolved.
I can’t be the first one to do this… right?
Just built a python script using one of @fofrAI’s models on @replicate to map out a grid of images of my face looking in every direction on my portfolio website, tracking the mouse on screen.
Should I drop the code on GitHub?
More work under the hood.
The #Velas team is gearing up for the next network upgrade, which will involve updates from the network validators.
🔥 Stay tuned to our channels, we'll keep you posted!
$VLX
🔥Lụm ngay airdrop khủng từ Mon Protocol nè ae 🔥
@monprotocol là dự án được back bởi các ông lớn trong ngành, bao gồm các ông lớn như Animoca Brands, Delphi Ventures. Đứng sau là team từng làm bộ sưu tập Pixelmon từng gây tiếng vang khi gọi vốn được hơn 70 triệu đô.
➡️Dự án đã xác nhận cho ae đổi point qua $MON khi token ra mắt. Các bước tham gia đơn giản như sau
1️⃣Vào link - Kết nối tài khoản X và điền code TuanHo17242 để nhận thêm 200 point
2️⃣Ae hoàn thành nhiệm vụ xong mỗi ngày sẽ được quay thêm 1 lượt.
3️⃣ Kiếm nhiều point sẽ có nhiều $MON nha ae
Anh em thấy hay thì cho mình 1 ❤️ và Retweet 🔁 để ủng hộ mình nha.
📌 Lưu ý : Kèo xịn cần làm sớm kiếm point nha ae
#AIRDROP #NICKY
[Kèo đầu cơ] Solana Mobile 2 có xứng đáng với $500?
Solana Mobile 2 đang được bán với giá $500. Cơ hội và rủi ro là gì?
1/ Solana Mobile (Saga) là gì?
2/ Lợi ích khi mua Solana Mobile 1 trước đây
3/ Lợi ích của Saga Mobile 2
4/ Rủi ro của Solana Mobile
5/ Cấu hình Saga Mobile 1 có tốt với tầm giá không?
6/ Tổng kết
Hi vọng các bạn ủng hộ mình link ref này https://t.co/oYZiI47M0X
Nếu threads chất lượng, mong nhận được ❤️ và 🔁 của các bạn 😍
#velas $vlx
Weekly updates:
- working on updating site content
- Desktop Wallets: updated all build versions, simplified build release, release itself in progress
- improved chrome extension code according the latest edits, now on review at Google
The First high performance Parallelized EVM Chain has hit a major milestone.
Sei’s permissionless Public Devnet has been upgraded to v2 and is now live for developers to begin their deployments.
It’s time to go even further beyond🔴
https://t.co/ZHtKVGxszI
There are only three ways to scale horizontally:
- Distributing
- Caching
- Asynchronous processing
But also many trade-offs.
Distributing
The concept is simple: spread the workload across many machines or systems.
Load balancing is the most common way of distribution. Sharding is a more complex one.
Pros:
Systems can handle more users and transactions without a single point of failure.
If one node fails, others can take over, enhancing the system's reliability.
Cons:
Managing a distributed system introduces deployment, networking, and data consistency complexity.
In the case of sharding. Data consistency across nodes is challenging, especially if you need strong consistency guarantees.
Caching
Storing copies of often accessed data in fast-access storage.
Pros:
It reduces response times by avoiding repetitive database queries or computations.
It decreases the load on the backend systems, allowing them to serve more requests.
Cons:
Data can get stale, and managing stale data is challenging.
Caching requires managing and scaling other resources (memory, cache servers).
Async Processing
Decoupling non-critical or intensive tasks using queues or event-driven architectures.
Pros:
Enhances the user experience by keeping the application responsive.
Enables better use of resources by spreading processing during off-peak times.
Cons:
Managing failures and data consistency in asynchronous operations is complex.
Delays in completing the background tasks can eventually affect consistency.
Remember, there are no easy solutions in distributed systems. It's just a collection of messy trade-offs.
Stay curious!
Self RAG
Self-reflection is one of the most interesting ideas in RAG, giving an LLM the ability to self-correct problems in retrieval and / or generation.
But, implementing this reliably can be tricky. We've found that graphs are a great way to reliably express logical flows, such as self-reflective RAG.
Here's a cookbook that shows how to implement ideas from the self-RAG paper by (@AkariAsai et al) using LangGraph.
It first uses a retrieval grader to (1) assess the quality of retrieved documents relative to the query. It then grades generations both for (2) hallucinations and for (3) usefulness in answering the question.
Cookbook:
https://t.co/3ThXzscI7g
Paper:
https://t.co/2oxTQOtViK
𝗪𝗵𝗮𝘁 𝗮𝗿𝗲 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿𝘀 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴?
O'Reilly recently shared their view on the tech trends, based on their internal learning platform metrics - from January through November 2022. and 2023.
Here are the main insights:
𝟭. 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁
There's a noticeable decline in software development in most topics, with 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 and design also experiencing a downturn. However, 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 saw an 8.9% increase, and event-driven architecture's usage, though small, rose by 40%. 𝗠𝗶𝗰𝗿𝗼𝘀𝗲𝗿𝘃𝗶𝗰𝗲𝘀 faced a 20% drop as developers debated their efficiency, preferring monoliths in some cases, while design patterns declined by 16%.
𝟮. 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲𝘀
Python remains the most popular, but 𝗝𝗮𝘃𝗮 and 𝗝𝗮𝘃𝗮𝗦𝗰𝗿𝗶𝗽𝘁 saw 14% and 3.9% declines, respectively. Interestingly, TypeScript's usage increased by 5.6%, suggesting a shift towards this language. 𝗖++ experienced a surprising 10% growth, and Rust's usage increased by 7.8%.
𝟯. 𝗔𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲
Natural language processing (NLP), generative models, and Transformers, saw dramatic increases in interest, with 𝗡𝗟𝗣 up by 195%, 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗺𝗼𝗱𝗲𝗹𝘀 by 900%, and 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗲𝗿𝘀 by 325%. Prompt engineering emerged as a new significant topic, paralleling Transformers in usage. 𝗗𝗲𝗲𝗽 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 and 𝗿𝗲𝗶𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 grew by 19% and 15%, respectively, while PyTorch and MLOps saw 25% and 14% increases.
𝟰. 𝗗𝗮𝘁𝗮
𝗗𝗮𝘁𝗮 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴, despite a slight decline of 3.6%, remains a heavily used topic, stabilizing after previous gains. 𝗞𝗮𝗳𝗸𝗮 and 𝗦𝗽𝗮𝗿𝗸 usage dropped significantly, while 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗣𝗼𝘄𝗲𝗿 𝗕𝗜 and 𝗦𝗤𝗟 𝗦𝗲𝗿𝘃𝗲𝗿 saw increases. Data warehouses experienced an 18% decline, but data mesh content usage rose 5.6%.
𝟱. 𝗗𝗲𝘃𝗢𝗽𝘀
𝗟𝗶𝗻𝘂𝘅's usage decreased by 6.9%, yet it remains dominant. 𝗞𝘂𝗯𝗲𝗿𝗻𝗲𝘁𝗲𝘀 and 𝗗𝗼𝗰𝗸𝗲𝗿's positions shifted slightly, with DevOps and SRE experiencing minor declines.
𝟲. 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆
Security topics generally grew, with 𝗻𝗲𝘁𝘄𝗼𝗿𝗸 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆, 𝗳𝗶𝗿𝗲𝘄𝗮𝗹𝗹𝘀, and 𝗵𝗮𝗿𝗱𝗲𝗻𝗶𝗻𝗴 seeing increases. 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 grew by 42%, highlighting a growing awareness among developers and operations staff.
𝟳. 𝗖𝗹𝗼𝘂𝗱 𝗖𝗼𝗺𝗽𝘂𝘁𝗶𝗻𝗴
Cloud computing trends show a 175% increase in 𝗖𝗹𝗼𝘂𝗱 𝗡𝗮𝘁𝗶𝘃𝗲, indicating a shift towards cloud-first development. 𝗖𝗹𝗼𝘂𝗱 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 and 𝗶𝗱𝗲𝗻𝘁𝗶𝘁𝘆 𝗮𝗻𝗱 𝗮𝗰𝗰𝗲𝘀𝘀 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 (IAM) also grew, while hybrid cloud saw a significant 145% increase. 𝗔𝗪𝗦 had a marginal gain, but 𝗔𝘇𝘂𝗿𝗲 and 𝗚𝗼𝗼𝗴𝗹𝗲 𝗖𝗹𝗼𝘂𝗱 faced notable declines.
𝟴. 𝗪𝗲𝗯 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁
Web development trends are mixed, with 𝗥𝗲𝗮𝗰𝘁 maintaining its position despite a slight decrease. 𝗔𝗻𝗴𝘂𝗹𝗮𝗿 and 𝗩𝘂𝗲 experienced changes in usage, with Vue showing a promising 28% growth. 𝗗𝗷𝗮𝗻𝗴𝗼 and 𝗣𝗛𝗣 both saw increases, while 𝗪𝗲𝗯𝗔𝘀𝘀𝗲𝗺𝗯𝗹𝘆 and 𝗕𝗹𝗮𝘇𝗼𝗿 faced declines, suggesting evolving preferences in web development technologies.
Check the full report in the comments.
#technology #softwareengineering #programming #techworldwithmilan #trends
$VLX | @VelasBlockchain
Nice uptrend. We are back at support. If you are looking for a entry this would be a good level.
Creating a nice weekly reversal candle.
If we reclaim the 0.03375 this beast will fly hard.
#Velas#vlxusdt#Layer1 $vlx
@ervichov@SFarkhad@hedera
The BRC-420 is the first Metaverse Protocol in Ordinals, enabling inscription assets to function across engines, games, and chains.
It also grants inscriptions a 'royalty' feature, allowing creators receive royalties from usage or recursions.
https://t.co/pSUhS5h8CY
1/4
👋 Hey Velas community, we think it would be amazing for all parties to add support to #Velas EVM Chain on @DeBankDeFi and @Rabby_io
🦾That's why we're here to explain how you can contribute.
It will take you 3 minute and must be done from a computer.
Read below 👇