Blog
The system doesn’t reward seam-fixers
Frontend Thinking, Responsibility Inversion, and Pay–Reward Asymmetry
AKA, Walking Along Component Boundaries: Why the Computer Industry Punishes People Who Don't Reinvent the Wheel
AKA, Why Open Source Is Getting Hit by AI Contributions
Walking Along Component Boundaries
The 201 PRs I’ve submitted to other people’s projects (click to expand)
Mojo-Webqq #255、Mojo-Webqq #261、Mojo-Webqq #276、docker #6、Install-Lion #2、QvPlugin-Command #5、Magisk #4613、ATLauncher #507、ATLauncher #511、Android-Architecture-knowledge-2- #4、bilimiao2 #8、magisk-unifiednlp-aosp-enabler #2、download-navi #263、bilimiao2 #17、bilimiao2 #18、NewPipeExtractor #936、LibreTube #3624、LibreTube #3632、linux-enable-ir-emitter #104、oiledmachine-overlay #81、distrobox #886、AM #58、AM #59、ansible-runner #1306、restic-android #199、bilimiao2 #67、QLDependency #50、Get_JDCookie #5、config #210、CachyOS-kernels #13、CachyOS-kernels #14、grpc-rust #1692、displaycal-py3 #380、distrobox #1455、eselect-llvm #3、stickerpicker #78、bilimiao2 #160、xarblu-overlay #507、InfiniTime #2143、distrobox #1622、hev-socks5-tunnel #208、hev-socks5-tunnel #219、LibreChat #7584、AM #1589、NCM-Downloader #14、installkernel-gentoo #36、GradleAndroidRustPlugin #10、GradleAndroidRustPlugin #11、kernel-patches #132、kvm-guest-drivers-windows #1473、kvm-guest-drivers-windows #1474、kvm-guest-drivers-windows #1475、kvm-guest-drivers-windows #1479、virtio-win-guest-tools-installer #85、kernel-patches #135、distrobox #1982、distrobox #1987、virtio-win-guest-tools-installer #87、virtio-win-guest-tools-installer #88、GradleAndroidRustPlugin #14、zram-init #57、zram-init #59、rclone #1、kvm-guest-drivers-windows #1536、kvm-guest-drivers-windows #1537、SillyTavern #5309、SillyTavern #5316、SillyTavern #5333、SillyTavern #5334、AM #2176、AM #2177、SillyTavern #5351、python-evdev #251、freebsd-src #2116、guru #1、rclone #2、reframe #25、reframe #26、hermes-webui #2189、hermes-webui #2190、honcho #689、opencode-models-discovery #3、plannotator #805、OmniRoute #2975、OmniRoute #3064、plannotator #843、OmniRoute #3094、OmniRoute #3146、OmniRoute #3188、OmniRoute #3422、OmniRoute #4293、OmniRoute #4359、neomacs #134、neomacs #135、neomacs #136、neomacs #137、neomacs #138、pi-hindsight #444、pi-honcho-memory #16、pi-mcp-adapter #159、OmniRoute #6322、OmniRoute #6323、OmniRoute #6330、OmniRoute #6336、OmniRoute #6702、OmniRoute #6714、OmniRoute #6727、pi-subagents #136、pi-subagents #139、OmniRoute #7004、OmniRoute #7012、OmniRoute #7027、pi-mcp-adapter #181、pi #6697、pi #6764、pi-subagents #153、pi-subagents #154、OmniRoute #7673、OmniRoute #7700、OmniRoute #7747、OmniRoute #7776、OmniRoute #7843、pi-subagents #169、OmniRoute #8367、pi-subagents #173、pi-tasks #42、pi-subagents #180、OmniRoute #8908、pi-hermes-memory #140、pi-subagents #191、pi-hermes-memory #141、pi-hermes-memory #142、OmniRoute #9199、OmniRoute #9205、pi-subagents #198、OmniRoute #9262、OmniRoute #9290、OmniRoute #9313、OmniRoute #9329、pi-intercom #87、OmniRoute #9555、OmniRoute #9569、pi-subagents #839、pi-hermes-memory #165、OmniRoute #9788、pi-subagents #922、OmniRoute #9983、OmniRoute #10066、codebase-memory-mcp #1530、OmniRoute #10072、pi-honcho-memory #21、OmniRoute #10079、pi-subagents #973、OmniRoute #10123、OmniRoute #10125、OmniRoute #10127、OmniRoute #10162、pi-caveman #15、OmniRoute #10243、OmniRoute #10247、pi-subagents #1141、pi-hermes-memory #189、pi-subagents #1214、rpiv-mono #174、OmniRoute #10712、OmniRoute #10716、OmniRoute #10723、OmniRoute #10806、deno #36644、pi-subagents #1297、pi-hermes-memory #191、pi-hermes-memory #192、OmniRoute #10953、OmniRoute #10963、sub2api #6139、CLIProxyAPI #5246、CLIProxyAPI #5247、CLIProxyAPI #5264、EternalTerminal #1、pi-subagents #1747、EternalTerminal #803、wenai #1、tproxy-config #36、sub2api #7215、sub2api #7217、rpiv-mono #241、honk #5、neomacs #412、rpiv-mono #262、CLIProxyAPI #6189、gpt-with-taste #6、sub2api #8191、cpa-quota-estimator #4、CLIProxyAPI #6760、sub2api #8302、aisix #1370、aisix #1387、youki #3862、youki #3866、youki #3867、honk #450。
I love the feeling of fixing the things I use every day and making them better for everyone, and I always assumed that if I did enough, I’d get more in return—make more money, perhaps, or become better known in the open-source community and land a better job. But unfortunately, I’ve gradually come to realize that my labor has become a stepping stone for others, while my name fades away without a sound.
To me, as someone who loves open source, it feels as though the open-source community is hurting each other. The more one person does for compatibility, edge cases, and the features they themselves need, the more of a failure they become—while the project owner becomes all the more successful. I don’t doubt others’ success, but I can’t help reflecting: am I doing this only for a humanity that doesn’t include me?
Eries' Trivia · Part Two
Author: Eries
Only when reading rather old books on performance optimization
when trying to reproduce their examples and looking at the assembly
and finding that everything has become utterly different
only then can one truly feel
the advances of instruction sets/compilersEries' Trivia · Part One
Author: Eries
Just now, while enslaving AI to work, I finished reading the short story "The Discrete Charm of the Turing Machine"
I started thinking about what kind of work I could have AI do to make a living for me
After much thought, it turns out there is only bounded-program work
My life seems to have reached its endViewing an LLM as an Interpreter: Managing Context Properly — Prompts, Memory, Function Calls, Skills, and Subagents
Humanity’s problem is that the people doing the work are always too young. The AI industry is rediscovering Cybernetics (Wiener, 1948) [1].
Do not imagine “intelligence,” but do not think of it as merely mechanical either.
The Last Dance Under the AI Bubble
I started buying gold after Silicon Valley Bank collapsed on March 10, 2023. The FDIC later confirmed that SVB was closed by California regulators on March 10, 2023, and announced on March 13 that it would protect all depositors; the FDIC’s deposit insurance reform report that same year also placed this event in the context of re-examining the U.S. deposit insurance system. [G1][G2] Since then, gold has tripled.
Gold is about to return to its essence and once again challenge the authority of the dollar.
Today, June 12, 2026, SpaceX is listed on Nasdaq. [X5] The IPO that tests the market’s temperature has arrived.
OWASP LLM 2025 Top 10 Study Notes
This is a supplement to the previous article AI Agent Privacy and Protection.
That article mainly covered AI Agent privacy, tool invocation, prompt injection, and subagent isolation. Here, referencing the OWASP Top 10 for LLM Applications 2025, I fill in the gaps around system prompts, training data, memory systems, toolchains, cost attacks, and side channels that were not expanded on before.
Since I'm currently focused on building AI Agent systems themselves, I won't discuss the security risks of training AI here for now.
The most common mistake in LLM security is treating AI as a "smarter backend service." But it's more like a probabilistic interface that is influenced by context, can call tools, can save memories, can incur costs, and can even be reverse-observed.
The value of the OWASP LLM 2025 Top 10 is not to give us a checklist to "tick off and be secure," but to remind us that the attack surface of LLM applications is no longer just the model itself, but the entire system composed of model, context, tools, data, memory, permissions, costs, and network traffic.
AI Agent Privacy and Protection
How to build an AI Agent that protects your privacy and your data against prompt injection
First, to analyze the security issues of AI Agents, we need to return to the essence of LLMs: An LLM is not an Agent; it is a model that generates tokens based on context.
Why Ultimately Younger People Become More Proficient with AI Than Older Generations
I found that I cannot use AI to learn/solve philosophical thinking.
Professionals can easily solve problems that I couldn't resolve after years of thinking and reading.
Reflecting on myself, perhaps the AI I use is only searching within "my" limited vector space.This led me to wonder, what kind of people truly "know how to use" AI?
Older people use new tools to solve existing problems, such as work issues; while children/students use AI to address academic problems.
Often, we assume that compared to "adult" problems like work, the academic challenges faced by younger people are simpler and more systematic. However, in reality, younger people don't have time to learn the meta-problem of "how to learn"—here, the "meta-problem" refers to their lack of opportunity to develop the habit of following the "traditional path." Consequently, they have ample motivation to seek their preferred "shortest path," which is inevitably validated afterward by assessments like "exams."
People Who Keep Watch While AI Works
People who keep watch while AI works close their eyes, and their thoughts gradually drift away.
People who keep watch while AI works,
close their eyes, and their thoughts gradually drift away.He thinks back through the ages—
the farmers who kept watch over rice paddies and reapers,
the workers who kept watch over lathes and robotic arms.Optimistic and Fail-able Queues in the AI Era
This article stems from reflections while fixing concurrent write issues for MCP's official Memory Server PR #3286.
Background: AI frequently calls MCP for memory storage, and simultaneous writes to the same file by multiple AI clients can directly corrupt the file format.
Choosing a Solution
The obvious choice: pessimistic locking.