Hacker News: jFriedensreich

New comment by jFriedensreich in "Show HN: Stage – Putting humans back in control of code review"

jFriedensreich — Fri, 17 Apr 2026 16:14:27 +0000

Looks kind of neat like devon.ai review / reviewstack crossover. But as i tell every of the dozens projects trying to make a commercial review tool: i would rather spend a week vibe copying this than onboarding a tool i have to pay for and am at the mercy of whoever made it. Its just over for selling saas tools like this. For agents i also need this local not on someones cloud. Its just a matter of time until someone does it.

New comment by jFriedensreich in "The beginning of scarcity in AI"

jFriedensreich — Fri, 17 Apr 2026 15:55:06 +0000

This is probably even the "fun" part of the whole picture. The purely dystopia starts when investment firms just silently grow bigger and bigger data centers like cancer. There will be no press releases, no papers, no chance anyone without billions will even know the details yet alone get access. One day we realise the worlds resources (maybe not as in the paperclip maximiser, but as in memory, energy, GPUs, water, locations) are consumed by trading models and the data centres are already guarded by robot armies. While we were distracted frighting with anthropic and openAI the real war was already over. Mythos is one sign in this direction but i also met a few people who were claiming to fund fairly large research and training operations just by internal models working on financial markets. I have no way to verify those claims but this happened 3 times now and the papers/research they were working on looked pretty solid and did not seem like they were running kimi openclaw on polymarket but actual models on some significant funds. Would be really interested if anyone here has some details on this reality. I would also not be surprised if this is a thing that people in SF just claim to sound dangerous and powerful.

New comment by jFriedensreich in "Building a CLI for All of Cloudflare"

jFriedensreich — Mon, 13 Apr 2026 21:16:56 +0000

I really dislike wrangler, though i understand the need to make something for mainstream appeal (agent cli tools + npx execution) It would be really nice if the cli was a layer on top of a git ops enabled declarative layer that was usable directly. (The terraform provider does not count...)

New comment by jFriedensreich in "Artemis II is competency porn"

jFriedensreich — Sun, 12 Apr 2026 11:04:14 +0000

In spite? You have yo be kidding. They sure showed him by giving him what he wants including public phone call and progress towards his moon plans that he talked about since 2017.

New comment by jFriedensreich in "I still prefer MCP over skills"

jFriedensreich — Sat, 11 Apr 2026 13:11:45 +0000

You assume an MCP has to work a certain way that is not the case. MCP can work however you want, its just a protocol. The same answer applies to tools as applies to skills. A tool has to look exactly the same to the LLM no matter if its seved from a cli or an MCP or a js function framework level tool. Credentials have to be injected in the gateway in either case.

New comment by jFriedensreich in "Google open-sources experimental agent orchestration testbed Scion"

jFriedensreich — Sat, 11 Apr 2026 13:06:28 +0000

Anyone at the frontier is switching to jj. Btw your question is kind of offensive, as if there is a universal truth "who matters" and everyone else can be dismissed. Companies do not matter for sure, if that was your premise.

New comment by jFriedensreich in "NanoClaw's architecture is a masterclass in doing less"

jFriedensreich — Sat, 11 Apr 2026 13:01:44 +0000

There are tons that are all too simple or too complex or made for mesh networking in k8, or too slow and made for debugging. My go-to solution is to make ad-hoc proxies with workerd that work just the way i need and compile to a "binary". It used to be annoying to write capnp by hand but now you just point an agent at worker.capnp and tell it what you need. The output is concise enough to be nice for proper reviewing and as extendable starting point, so neither vibe coding nor much manual coding needed.

New comment by jFriedensreich in "Artemis II is competency porn"

jFriedensreich — Sat, 11 Apr 2026 12:00:57 +0000

As a german, seeing you go to the moon under trump feels like celebrating the olympics in the dawn of the german reich, it cannot be taken for what it is. It does not matter what we feel, what we want to feel, we can enjoy it for a second and then swallow it down and not write an article like that.

New comment by jFriedensreich in "OpenClaw’s memory is unreliable, and you don’t know when it will break"

jFriedensreich — Sat, 11 Apr 2026 11:38:48 +0000

Memory systems as most people understand and build them are a clear dead end. We just need skills, tools and better context management.

New comment by jFriedensreich in "I still prefer MCP over skills"

jFriedensreich — Fri, 10 Apr 2026 19:16:44 +0000

Don't think its relevant who they are if they give advice that is based on outdated understanding of how agent harnesses are build and how to use MCP in an agent harness in the first place. You can serve agent skills via mcp or via text files accessed via local tools, if your harness makes this look different to the LLM in the end it is just a bad harness. The LLM should just see "ways to discover skills" and then "use the skills". If skills come from a folder or from an MCP is transparent implementation detail. This is more than just theoretical, if abstractions like the way skills are served leak into the context, this will measurably degrade agent performance, depending on model more or les severe!

New comment by jFriedensreich in "I still prefer MCP over skills"

jFriedensreich — Fri, 10 Apr 2026 16:02:54 +0000

If your llm sees even a difference between local skill and remote MCP thats a leak in your abstraction and shortcoming of the agent harness and should not influence the decision how we need to build these system for the devs and end users. They way this comment thinks about building for agents would lead to a hellscape.

New comment by jFriedensreich in "Dropping Cloudflare for Bunny.net"

jFriedensreich — Wed, 08 Apr 2026 21:08:27 +0000

Maybe i misunderstood what you were going for. Just saying: Yes, their web fetch handler is not a web standard but the only standard there is to accomplish something like that had a few issues and they picked something that looks as generic as possible. vs import {pretendToStartServer} from ('our:global-injected-pseudo-sdk')

I really loved the idea of using service worker sdk and having something that works in browsers too so the conceptual model of edge functions are spanning backends, edge POPs and browser runtimes.

New comment by jFriedensreich in "Dropping Cloudflare for Bunny.net"

jFriedensreich — Wed, 08 Apr 2026 14:01:51 +0000

but not to lock people in but to avoid the context overhead of service worker api, which they tried first and which still works btw.

New comment by jFriedensreich in "Dropping Cloudflare for Bunny.net"

jFriedensreich — Wed, 08 Apr 2026 13:49:57 +0000

Have you seen the codebase? Please don’t blanket reply like that just because it is a corp. Workerd is one of my favorite piece of engineering and its not comparable at all to openwashing projects that have extreme internal garbage not possible to free from the companies infrastructure. Kenton and others fought hard to make the abstractions among the cleanest and well designed in the industry and while it is not a multi contributor apache foundation style project it is a codebase that you could hard fork if cf decided to close it or go a direction you cannot accept.

New comment by jFriedensreich in "Google open-sources experimental agent orchestration testbed Scion"

jFriedensreich — Tue, 07 Apr 2026 19:43:57 +0000

Disapointing google of all places uses git worktrees instead of jj workspaces.

New comment by jFriedensreich in "Project Glasswing: Securing critical software for the AI era"

jFriedensreich — Tue, 07 Apr 2026 19:25:53 +0000

The only thing reassuring is the Apache and Linux foundation setups. Lets hope this is not just an appeasing mention but more fundamental. If there are really models too dangerous to release to the public, companies like oracle, amazon and microsoft would absolutely use this exclusive power to not just fix their holes but to damage their competitors.

New comment by jFriedensreich in "Dropping Cloudflare for Bunny.net"

jFriedensreich — Tue, 07 Apr 2026 17:16:05 +0000

The platform yes, but the runtime and the APIs are apache 2.0 licensed and uses web standards wherever possible. Thats way more open than most alternatives i am aware of.

New comment by jFriedensreich in "Dropping Cloudflare for Bunny.net"

jFriedensreich — Tue, 07 Apr 2026 16:32:37 +0000

Because an edge runtime should not directly start servers in the first place. You provide handlers that fulfil requests from the system. Edge runtimes that pretend to start the servers in the edge worker context are flawed and ugly from the very start. APIs like that leak implementation details and increase vendor lock-in for no added benefit.

New comment by jFriedensreich in "Dropping Cloudflare for Bunny.net"

jFriedensreich — Tue, 07 Apr 2026 15:52:46 +0000

Cloudflare is not a CDN anymore but the workers edge platform, if you can move to bunny.net, you were not really using cloudflare. I don't understand how none of the alternatives really embrace WinterTC

If i see something horrific like:

import * as BunnySDK from "@bunny.net/edgescript-sdk" BunnySDK.net.http.serve(async (request: Request) =>

Thats a proprietary lock-in worse than what it tries to replace!

New comment by jFriedensreich in "NanoClaw's architecture is a masterclass in doing less"

jFriedensreich — Tue, 07 Apr 2026 14:31:07 +0000

The "permissions as access/visibility" is unfortunately not holding up in practice. As nice as a simplification like this would be: No one wants to configure up front what permissions are needed. Users will run into annoyances and then just overcommit resource access the same way they choose to run yolo mode. Limiting access for security is a great part of a capability based system but a system that will be used in the way intended needs the following properties, which are mostly missing from the tools we have (apparmor, seccomp, oss VMs, seatbelt, docker, bubblewrap etc.)

- pause execution for a policy engine or user input

- variable scope permissions independent of what was requested. eg user needs to allow just this request to /test/myfile.txt or grant /test/* in that moment

- add (ideally also remove) capabilities based on dynamic user input or engine decisions without up front configuration

- not need application support, if apps need to support it the moment the harness uses an external tool the model breaks

deno, workerd and maybe a vm/docker solution with an webdav proxy mount and web-proxy are the only environments i am aware of where systems like this could be build at all, even there, with limitations. (Not writing this to sound absolute but to learn about other options I am missing.)