Founder of Baixing.com: My Experience with Claude Code in Fourteen Points

By: rootdata|2026/06/09 03:45:01

Author: Wang Jianshuo

A simple record of my experiences with Claude Code so far, purely personal exploration, not necessarily applicable to everyone.

Focus on one tool and use it intensively. I use Claude Code. I don't think it's better than Codex, but the ROI of comparing tools may not be high, even though it can create a false sense of achievement by articulating the differences convincingly.
Remember the most important shortcuts. Control+G opens the editor, which helps in writing longer content; Control+A, Control+E, Control+U are very useful shortcuts for quickly moving the cursor in the command line. Although they are not new in the AI era, they are as important as Control+C and Control+V when in use.
Use voice input. HoldSpeak is very helpful.
For a project, first write PROJECT.md, using a structured approach to jot down everything you think of in one go.
Claude agents are the default opening method.
Claude Code, github.com, and cloudflare.com are a perfect match, handing over all operations related to the build process, release process, and domain management to the infrastructure.
Separate what is written by humans and what is written by machines. Manually maintain the core CLAUDE.md, and do not read the .md or code written by Claude Code. Machines belong to machines, and humans belong to humans. Understand what AI writes by asking AI, not by looking at the source code.
Drag files into the Claude Code window—audio, video, documents, screenshots—if it's unclear, use Command+Shift+5 to take a screenshot and then drag it over, the fastest way.
Reconstruct the memory system. Centered around ~/.claude/CLAUDE.md, categorize and reference multiple memory files, requiring not to use the project's memory, and place all memory files in git, syncing to GitHub (private), so that your memory is permanent and accumulative, rather than scattered across each project.
Write Skills, and after each work session, ask Claude to "consolidate what has been learned into Skills"—it can do this automatically.
If possible, use ultracode to trigger dynamic workflows for complex tasks. Although it's expensive and slow, the results are guaranteed.
Accumulate skills along the way, and continuously reconstruct skills. Skills need to be stored in git.
Use the documentation from git as the output of the previous task and the input for the next task. Ensure there are clear handover documents between agents, rather than relying on context for continuity.
Treat Claude Code as a horse (or a person), not as a vehicle. A vehicle turns under your command, while a horse has its own ideas; we just need to set goals and boundaries. Its autonomous pathfinding feature is a characteristic, not a bug.

Does anyone have anything to add?

Solana did not fall behind during the bear market. Trading enthusiasm has waned, but the network is more stable, RWA and stablecoins are expanding, and the capital foundation is much thicker than in the previous cycle. The real question is: when the speculative tide recedes, can perpetuals, predicti...

Young people in South Korea make a "final effort" in the epic bull market

The South Koreans' average of two accounts for wildly gambling in the chip bull market reflects the survival anxiety and harsh reality of countless young people trying to break through class barriers behind the nationwide stock trading frenzy for wealth.

Dialogue with OmenX Founder: Why does the prediction market need an evolution from "spot" to "derivatives"?

How to reconstruct the prediction market using leverage?

When the P2P illicit funds from ten years ago turned into 60,000 bitcoins

The largest Bitcoin money laundering case in the UK has new developments: 16,000 Chinese victims are pursuing 61,000 seized Bitcoins across borders, and the dispute over the applicability of UK and Chinese laws will directly determine whether the victims can share in the soaring profits.

Morning News | CME Group launches Nasdaq Cryptocurrency Index futures; Asset management giant Janus Henderson strategically invests in Ethena

Overview of Important Market Events on June 10

Why did Oracle deliver the strongest financial report in history, yet its stock price fell?

Oracle's revenue for fiscal year 2026 set a record, with AI cloud orders soaring to $638 billion, but massive capital expenditures on computing power led to negative free cash flow, causing a 5% drop in after-hours stock prices.

Bitcoin Layer 2 Network Botanix: Why Did We Choose to Dissolve?

The Bitcoin L2 star project Botanix announced a gradual shutdown, with the team admitting to facing severe challenges from the failure of its business model and the prevailing trends. Users are urged to withdraw all assets before July 9, 2026.

Morning Report | OpenAI has submitted an S-1 registration statement draft to the U.S. SEC; Morpho completes $175 million financing

Overview of Important Market Events on June 9th

Galaxy Deep Research Report: How Hyperliquid's HIP-4 Upgrade Changes the Landscape of Prediction Markets?

The platform that wins this competition will be the one whose execution layer is the hardest to replicate, whose builder ecosystem delivers the fastest, and whose regulatory path is the most open.

Latest research from 13 top universities including Cornell University: The current state, challenges, and misconceptions of the fusion of Crypto and AI

The combination of AI and crypto is still in its early stages, with both serving as complementary "middleware": AI translates human intentions into executable programs, while cryptographic technology provides verifiable and tamper-proof guarantees for computational processes and results. In the dire...

Deconstructing Anthropic: The Best AI Company, Possibly Also a Type of Organizational Invention

Instead of competing with ambition, focusing on restraint, how does Anthropic leverage extreme strategic focus and an "counterintuitive" geek culture to counterattack OpenAI on the AI battlefield?

Every exchange is a "Universal Exchange."

You initially build infrastructure for something, then realize it can also be used for many other things, and then you continuously expand the business to accommodate everything that the infrastructure can support.

The counterattack of traditional finance: Alliance chains are quietly reviving

Whether public chains win or consortium chains win has never been the focus.

Pantera Capital Partner: How Tokenization is Restructuring the Private Equity and Early Investment Ecosystem?

Top tech companies are going public later and later, leaving retail investors shut out during the high growth period. Can tokenization give ordinary people back this entry ticket?

Mastercard Launches Agent Pay for AI, Plans to Record AI Agent Payment Authorizations on Polygon

Mastercard launched Agent Pay for AI, a new payment protocol designed to help AI agents make small payments such as pay-per-use access to data and APIs. The system plans to record human-granted AI agent permissions on Polygon, focusing on verifiable authorization, identity, and payment controls.

Curve Deploys Llamalend v2 on Optimism With 250,000 OP Incentives

Curve launched Llamalend v2 on Optimism with 250,000 OP incentives from the Optimism Foundation. The upgrade expands Llamalend beyond its earlier crvUSD-focused model, adding broader collateral support, LlamaRisk market reviews, and the ability to use Curve LP tokens as collateral.