HNHacker News
TopNewBestAskShowJobs

pbronez

2,117 karma · joined May 10, 2019

[ my public key: https://keybase.io/peterbronez; my proof: https://keybase.io/peterbronez/sigs/aK3o2l59iuXf4ZE61-arwlc3CsPE1jYuB6pu_vjfi0c ]

<a rel="me" href="https://hachyderm.io/@PeterBronez">Mastodon</a>

submissionscomments
pbronez··on Incident Report: Railway Blocked by Google Cloud (Resolved)
Yeah I’m planning a similar transition for my personal infrastructure. Railway is super easy to get started with, dashboard and logs features are nice, but I’ve just lost confidence in it.
pbronez··on Hindenburg’s Smoking Room
This is a really lovely website. I had no idea that stamp collectors were a critical market for airships!

> The polar journey, like other zeppelin flights, was largely financed by stamp collectors; Graf Zeppelin carried approximately 50,000 letters sent by philatelists, and made a water-landing to exchange mail with the Soviet icebreaker Malygin, which itself carried a large quantity of mail sent by stamp collectors.

pbronez··on δ-mem: Efficient Online Memory for Large Language Models
Interesting approach.

> Prioritize recall over precision.

Have you tried stemming your regex? That would help you catch messages where a different form of your word appeared. For example instead of “story” you look for “stor” which catches “stories” as well.

Then you might think, could we do an even better job by figuring out the general semantic intent of the query and history? Let’s project them into a semantic vector space! That’s an embedding.

Then you want to query that, which means you need a vector database. So now we can take the query, embed it, query the vector DB with that embedding and retrieve the N closest history documents. You can use that to augment the generation of the response to your prompt.

This is RAG.

Anyway, interesting to see different degrees of sophistication here. Certainly a handful of naive regex are very snappy.

There’s probably a hybrid approach where you use sophisticated NLP and embedding techniques to robustly define topics, then train a regex to approximate that well.

pbronez··on Show HN: Find the best local LLM for your hardware, ranked by benchmarks
Cool, but it looks like it doesn’t actually test anything on your machine? It does hardware detection and then some lookups. Maybe I missed it but I really want a tool like this to actually run a model on my machine to get the speed numbers.

I’ve been using RapidMLX for this. The integrated speed tests matter because the quality of the backend is a moving target and the quantization / MLX format conversion also matter. It’s not enough to say “oh use this model family with X parameters” you have to add the architecture specific quantization too.

https://github.com/raullenchai/Rapid-MLX

pbronez··on Twin brothers wipe 96 government databases minutes after being fired
> We wanted to release a Windows version as part of Windows 98, but sadly, Microsoft has effective building security.
pbronez··on Twin brothers wipe 96 government databases minutes after being fired
That’s the “firming up” bit. You have a contract that deems the code “work for hire” even though the money flow is wonky. Legally the guy is like any 1099.
pbronez··on Our Continuation of MkDocs
Yeah that’s a better link for the overall situation. Personally I’m betting on Zensical.
pbronez··on Yet Another GitHub Incident
This video suggests OpenAI is actively developing an alternative to GitHub

https://youtu.be/f3u57jkwBFE?si=FJuxZfmc-i7EkPlx

pbronez··on PyInfra 3.8.0 Is Out
How does this compare to Salt Stack?

“Built on Python, Salt is an event-driven automation tool and framework to deploy, configure, and manage complex IT systems. Use Salt to automate common infrastructure administration tasks and ensure that all the components of your infrastructure are operating in a consistent desired state.”

https://docs.saltproject.io/en/latest/topics/about_salt_proj...

pbronez··on Apple accidentally left Claude.md files Apple Support app
I think it’s more like project-wide code style rules or build instructions.
pbronez··on Mozilla's opposition to Chrome's Prompt API
I’m kinda terrified by the security implications of the Prompt API.

This is a way for web services to make your computer complete large amounts of compute at their behest. Tokens have value. There will be incentive for bad actors to use your local LLM for their own purposes, much like hostile crypto mining payloads.

This is an obvious target for prompt injection attacks and other malicious remote code execution. In many ways, model prompts ARE programs. The browser / local device would need to provide an LLM with the same sandbox guarantees as the rest of the browser. Can they be trusted to do that? Does anyone understand this well enough to do that with confidence?

I’m a big fan of local models, but I would be very cautious about letting random websites call the model I’m hosting on my local machine with open source software.

pbronez··on Mozilla's opposition to Chrome's Prompt API
Yeah I looked into the Apple Foundation models and was surprised at their limited scope. On reflection it made sense though. They’re giving you the small part of the LLM capability surface that (1) can run with good performance on all their hardware and (2) works reliably.

It’s not enough for a chat-first research agent, but it’s definitely enough to unlock features that rely on natural language understanding. Seems like a small thing compared to Claude/ChatGPT and the general hype, but still magic in its own context.

pbronez··on Before GitHub
Maybe something like https://fossil-scm.org/home/doc/trunk/www/index.wiki is worth another look. Although Forgejo is probably an easier switch from GitHub. If their federation ideas play out well, that could be a good outcome from all this
pbronez··on An AI agent deleted our production database. The agent's confession is below
Railway also doesn't let you download the "backups" out of their platform. You can restore the backup to that instance of that service and basically nowhere else.
pbronez··on An update on GitHub availability
perhaps https://isolveproblems.substack.com/p/how-microsoft-vaporize... ?
pbronez··on What's missing in the 'agentic' story: a well-defined user agent role
Yes. I think you can get agents to “Conscious competence” with a lot of well-designed oversight, direction and control. It works, but it’s fragile - nothing like the judgement needed to handle novel situations well.

https://en.wikipedia.org/wiki/Four_stages_of_competence

pbronez··on I am building a cloud
Neat service. Website doesn't provide enough information for me to trust any workloads to it. Not clear where the underlying infrastructure is, what security guarantees I get, etc.
pbronez··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
Rapid MLX team has done some interesting benchmarking that suggests Qwopus 27B is pretty solid. Their tool includes benchmarking features so you can evaluate your own setup.

They have a metric called Model-Harness Index:

MHI = 0.50 × ToolCalling + 0.30 × HumanEval + 0.20 × MMLU (scale 0-100)

https://github.com/raullenchai/Rapid-MLX

pbronez··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
I haven’t tried it yet, but Rapid MLX has a neat feature for automatic model switching. It runs a local model using Apple’s MLX framework, then “falls forward” to the cloud dynamically based on usage patterns:

> Smart Cloud Routing > > Large-context requests auto-route to a cloud LLM (GPT-5, Claude, etc.) when local prefill would be slow. Routing based on new tokens after cache hit. --cloud-model openai/gpt-5 --cloud-threshold 20000

https://github.com/raullenchai/Rapid-MLX

pbronez··on Show HN: SmallDocs - Markdown without the frustrations
Cool project. Heads up - there’s a commercial company with a very similar name that might decide to hassle you about it:

https://www.sdocs.com/

pbronez··on Cloudflare Email Service
What structural changes could we make to improve the situation?
pbronez··on Show HN: boringBar – a taskbar-style dock replacement for macOS
Agree that treating each major tab within a terminal emulator / browser/ etc as a nearly-first-class window would be helpful. Windows actually does this pretty well.
pbronez··on Show HN: A WYSIWYG word processor in Python
Check out Typst. It’s a markup language focused on print with HTML layout as a secondary target.
pbronez··on How NASA built Artemis II’s fault-tolerant computer
The Artemis computer handles way more flight functions than Apollo did. What are the practical benefits of that?

This electrify & integrate playbook has brought benefits to many industries, usually where better coordination unlocks efficiencies. Sometimes the smarts just add new failure modes and predatory vendor relationships. It’s showing up in space as more modular spacecraft, lower costs and more mission flexibility. But how is this playing out in manned space craft?

pbronez··on Show HN: Apfel – The free AI already on your Mac
Digging into this, found Apple’s release notes for the Foundation Model Service

https://developer.apple.com/documentation/Updates/Foundation...

They released an official python SDK in March 2026:

https://github.com/apple/python-apple-fm-sdk

pbronez··on The revenge of the data scientist
Yup, agree. “Evaluations” = Tests

Gets pretty meta when you’re evaluating a model which needs to evaluate the output of another agent… gotta pin things down to ground truth somewhere.

pbronez··on Building FireStriker: Making Civic Tech Free
Classic example of a problem we need to solve with policy instead of technology. The law isn’t even that hard:

1) Government authored/sponsored materials are not considered “published” until they are available in both human and machine readable digital formats at no cost to the reader.

2) Publications of such materials

2.1) must not keep a record of individual usage beyond that technically necessary to provide access to the publication

2.2) not withstanding 2.1, take reasonable steps to secure digital publications from abuse that creates an undue administrative burden

pbronez··on How BYD got EV chargers to work almost as fast as gas pumps
Maybe. Agree that zero-sum thinking sucks. You gotta grow the pie. But. You also have to share the big pie.

In your example, the current crisis can be represented as:

A field exists and produces 1 unit.

A financial entity buys the field and applies unsustainable methods to increase production 100 units, keep 99.5 of them, distributes 0.5/n. People are pissed that they’re getting half of what they used to despite incredible productivity. The people elect a leader to fix the situation. The leader confronts the financial entity, and returns to the people with 4 units in their pocket and excuses.

pbronez··on Wanter – A tiny, decentralised tool to explore the small web
Cool idea - it’s like a recommended set of links, but integrated into an interface like stumble upon or Kagi’s small web browser
pbronez··on Quillx is an open standard for disclosing AI involvement in software projects
Neat idea. I like the five point scale
← PreviousPage 4 of 34Next →