🍒
Cherry Servers
We give Developers full control and flexibility to build cloud-native applications on a powerful Cloud Platform.
Hello Builder,
Here’s your weekly wrap-up. Solana’s Drift upgrade made on-chain trading much faster, and DeFi discussions showed why many users are turning to AI-driven tools to handle strategies that have become too complex to manage manually. The IMF called for stricter rules around stablecoins, and there were updates on new platform licenses.
Developers shared hands-on impressions of GPT-5.1, Gemini 3, and Opus 4.5, along with test results showing how these models perform in real coding and reasoning tasks. Teams also experimented with agents that can plan steps and use tools, and explored when Python is a better fit than shell scripts for automation.
Let's dive into the topics ⬇️:
🌐 Web3
Drift v3 makes its Solana-based derivatives DEX feel nearly as fast and smooth as a centralized exchange while remaining fully on-chain. By executing most orders within a single Solana slot and reducing slippage, it aims to make DeFi perpetuals practical for high-speed, active traders.
Web3 adoption in Europe is growing steadily but cautiously, supported by a mature crypto market, clearer regulations, and innovation in DeFi, NFTs. Its future will hinge on AI integration, real-world asset tokenization, and more structured DAOs.
Stablecoins challenge traditional finance's dominance, prompting IMF calls for harmonized standards, robust reserves transparency, capital controls, and CBDC prioritization to preserve sovereignty while harnessing payment efficiencies like low-cost remittances.
This introduces a high-performance, EVM-compatible blockchain that supports DeFi and high-frequency finance with fast transaction speeds. This level of scalability and compatibility enables developers to build complex applications without sacrificing speed or security.
DePIN like ProverNet decentralizes compute infrastructure, reducing costs, enhancing security via blockchain transparency, and enabling real-world scalability for ZK apps amid growing demand, fostering a shared ecosystem over silos.
📘 AI/ML
If you're into AI engineering, the latest Rate Limited episode is a standout recap of hands-on impressions of GPT 5.1, Gemini 3, and Opus 4.5, plus practical discussions on tooling like Cursor, Claude Code, Droid, terminal setups.
LangChain just shipped three updates: OpenAI-based content‑moderation middleware for agents, unified cost tracking that includes tools and custom APIs, and a DeepAgents CLI that scores roughly mid‑40s percent on Terminal Bench 2.0, comparable to Claude Code on that benchmark.
Recommended by LinkedIn
A Hugging Face user built “tool orchestrator” that lets any LLM emit small Rhai scripts instead of JSON tool calls. Those scripts run in a sandboxed Rust/WebAssembly runtime, which handles calling tools, looping, branching, and combining results without sending all intermediate data back through the model.
The idea here is that infrastructure is moving toward a model where AI handles the flexible reasoning, and a tightly controlled execution layer makes sure everything runs safely and predictably. FlowAI is Itential’s attempt to bring that approach into a real, usable product.
It supports extremely long text (131K tokens) for complex tasks but is text-only, unlike GPT-5 which supports multimodal inputs. DeepSeek excels in efficiency and cost-effectiveness, making it a strong alternative for many AI applications.
🔧 DevOps
OpenAI’s Codex coding agent, shows how it can act like a teammate that plans, writes, tests, and reviews code with you. It explains key concepts like AGENTS.md, PLANS.md, and GPT‑5.1 Codex Max so you can plug AI deeply into your everyday development workflow.
A prediction that DevOps will increasingly center on AI-driven automation, resilience, and stronger accountability across the delivery lifecycle. It also points to steady growth in platform engineering, integrated DevSecOps and observability.
The core idea is that once a script gets more complex than a one-liner, Python is usually a better tool than shell scripting. It’s more portable, easier to read and maintain, and behaves consistently across systems, making it the smarter choice for non-trivial automation.
Kubernetes now has a formal AI conformance program, created by the Cloud Native Computing Foundation (CNCF) to standardize how AI workloads run across different Kubernetes platforms, and VMware’s vSphere Kubernetes Service (VKS) is among the first certified implementations.
💡 Did you know?
The video shows how to run multiple Python trading bots 24/7 on a remote Cherry Servers VPS using VS Code/Cursor plus GNU screen, so bots keep running even when the local computer is off.
This tutorial explains how to install XRDP on Ubuntu 24.04 so you can manage your server through a secure, convenient graphical desktop from anywhere, with room for further customization and security tweaks.
