cahaya

I build my own take on this with https://document.bot Build for aviation legal regulations/ compliance. Highly optimized search around large PDF's, Word, etc with offline, EU and frontier LLM options.

Until recently my M1 Max 32GB 2021 was great until memory became the bottleneck. My M1 Pro 8GB lasted me less than a year bcs with Docker dev stuff made it sluggish. 32GB was great for years until AI assisted productivity got so high with Codex. Running 3 projects, 3 different worktrees, each either Docker compose or running/building Electron apps. So in this high AI era productivity my Mac is using insane amount of RAM and SSD space (bcs of all the worktrees)

Agree, with you but I'm still using 3.6 Flash because of tok/s/ latency/ uptime with high context. Tried Grok 4.6 and it was scoring lower on some internal benchmarks or slower.

Thanks for the post. For https://Document.bot (Kinda Cursor IDE for knowledge workers), im already trying to improve the harness (besides spotting bugs) using hillclimb experiments. More and more i'm using a AI harness engineering skill in Codex to further improve the app. This blog post helped me to improve the skill a bit.

Nice, @grok how does it compare to Cloudflare that also provides a REST endpoint for structured markdown data and screenshots?

When GPT-Live in Codex, so i can walk the dog while shipping?

I'm not disagreeing with the article, but I think you are right that if you use AI in the right setting it can do much more than the average ChatGP that outputs Linkedin like crap:

interdisciplinary teams/person + SOTA AI + Right context (customer/ internal pain points, access to the code, focussed Hackathon) = amazing cool things.

From X: Today, we are releasing Rampart: a 14.7MB machine learning model designed to protect citizens’ privacy by redacting personal information directly in your browser before it gets sent to any server.

Asking for feedback:

Sorry for hijacking the convo, but you (with local models) are my target audience in terms of hardware.

Is anybody willing to test my new app https://document.bot? It is like Cursor IDE but custom harness for knowledge work (PDF's, MS Office files etc).

You can connect your existing offline LLM models through LMStudio, Ollama, or app managed LLM models (Qwen3.5, Gemma 4, etc)

Might have to make a new Ask HN post for this, but again, you are users with good hardware setups.

Same. Solving it by moving complex and sensitive data to an offline desktop app https://document.bot that support offline (self hosted) AI models (and optionally EU/ US AI providers). However, it doesn't integrate yet with shared (org) drives.

Building Cursor IDE, but for knowledge workers. Domain will go live soon on https://document.bot

Hoping OpenAI/ Codex will launch this soon too.

Nice. Curious about 5.3-codex-high results

  Codex app-server is the interface Codex uses to power rich clients (for example, the Codex VS Code extension). Use it when you want a deep integration inside your own product.

It mentions 'Inside your own product', but not sure if that means also your own commercial application.

Same question here. A while ago I read rumors OpenAI might build a "Login with OpenAI" (comparable to login with Apple, Facebook, Google) so people can also use their existing sub in commercial apps. Hope it's true.

looks nice!

Lot's of OCR/ LLM's (Even Gemini Pro 2.5) still struggle converting complex tables to markdown or HTML: Tables with multiple headers and merged cells that get mixed up, multiple columns with tick boxes get mixed up, multi page tables that are not understood correctly. Also Llamaindex fails miserably on those things.

Curious to hear which OCR/ LLM excels with these specific issues? Example complex table: https://cdn.aviation.bot/complex-tables.zip

I can only parse this table correctly by first parsing the table headers manually into HTML as example output. However, it still mixes up tick boxes. Full table examples: https://www.easa.europa.eu/en/icao-compliance-checklist

Got 32GB so was hoping I could use ollm to offload it to my SSD. Slower but making it possible to run bigger models (in emergencies)

Nice. Seems like i cannot run this on my Apple silicon M chips right?

Nice, but can somebody tell me if this performs better than my simple Postgres MCP using npx? My current setup uses the LLM to search through my local Postgres in multiple steps. I guess this Pgmcp is doing multiple steps in the background and returns the final result to the LLM calling the MCP tool?

Codex: ``` [mcp_servers.postgresMCP] command = "npx" args = ["-y", "@modelcontextprotocol/server-postgres", "postgresql://user:password@localhost:5432/db"] ```

Cursor: ``` "postgresMCP": { "command": "npx", "args": [ "-y", "@modelcontextprotocol/server-postgres", "postgresql://user:password@localhost:5432/db" ] }, ```

With my setup i can easily switch between LLM's

Are there any existing scripts/ tools to use these evolutionary algorithms also at home with e.g. Codex/GPT-5 / Claude Code?

Is it me or is the link only opening as JSON?

{ "@context": [ "https://www.w3.org/ns/activitystreams", { "Hashtag": "as:Hashtag", "sensitive": "as:sensitive", "dcterms": "http://purl.org/dc/terms/" } ], "id": "https://neurofrontiers.blog/?p=11319", "type": "Note", "attachment": [ { "type": "Image", "url": "https://i0.wp.com/neurofrontiers.blog/wp-content/uploads/202...", "mediaType": "image/jpeg", "name": "A cartoon depicting a traffic jam of identical black cars, each driven by a brain. Thick clouds of exhaust rise and gather above the stalled vehicles." } ], "attributedTo": "https://neurofrontiers.blog/author/neuronerdb/", "audience": "https://neurofrontiers.blog/?author=0", "content": "<h2>How does air pollution

I tried adding Context7 Documentation MCP and got this

URL:https://mcp.context7.com/mcp Safety Scan: Passed

This MCP server can't be used by ChatGPT to search information because it doesn't implement our specification: search action not found https://platform.openai.com/docs/mcp#create-an-mcp-server

I was also looking for a video. The concept sounds good, but feels like I need to learn a lot of new commands, or have a cheat sheet next to me to be able to be able to use the framework.

Wondering how average users can benefit from this platform with Claude Code and the relation to Vending Bench that tracks how much money LLM's can make.

https://andonlabs.com/evals/vending-bench

Nice! Missing a cost calculator with input and output fields.

Exactly. If Conductor would work with local branches I would switch from Crystal.

I can confirm. When trying convert simple Word sentences and tables to e.g. Markdown/HTML from a Word XML you need a PhD in XML edge cases and nested garbage.

Building a AI-assisted aviation regulation compliance tool for aviation professionals: https://aviation.bot At the moment just a RAG with EU aviation regulations. Soon FAA, UK CAA, and more complex AI agent features that do more complex deep research.