AmazingTurtle

so basically dots is openclaw absorbed into chatgpt?

you can actually leverage 400k and 1M contexts in codex with very little code changes to the harness. note that excess context past the.. 250k or 400k mark (i don't remember) is charged at 2x the price.

regarding the new pro max subscription btw:

it's 500$ for 25x the plus usage, thats pro (max).

this implies that the old 200$ 20x pro (more) is now more like 10x the usage of plus.

they are slashing our subscriptions in half and make it "but we're more efficient!"

"Your scientists were so preoccupied with whether they could, they didn't stop to think if they should"

Had to be a team of circle jerk external advisories that want to become rich by inflating the effort here

Perfect, I will incorporate this as default as well as the commit from the other guy into my own codex fork https://github.com/AmazingTurtle/codex btw. I'm rebasing on 0.156.0 right now

My calculation (that is 3x 20x subscriptions) it will pay off after ~18.5 months, if I were to stop my subscriptions today. I will likely keep at least one though, so it's more like 27.5 months for a payoff. I am not doing it to save money though. I am doing it for security of supply. Constant quality - I know my model isn't getting lobotomized etc. - and open weight models mostly just lack very little behind. I'm sure I will have affordable, fast and efficient sol 5.6 capabilities with open weight models on my sparks within 12-18 months easily.

> Could there be a benefit to releasing a new model, slowly dumbing it down over a couple months, then releasing a new model that’s marginally if at all better than the original to create a perceived improvement when in reality there isn’t really one?

Exactly what I am saying for months now. And it's exactly the reason why I am shifting to open weight models now. Just bought myself a 2x DGX Spark Cluster. Will run Qwen3.8 Flash Next on it, maybe Qwen4 when it comes out.

Not only do I have full control over quantization and inference, but also will I experience a constant level of quality. It won't be frontier. But it will be stable, and that's enough reason for me to switch. Also I will likely save some money on subscriptions.

This is the first time that I hope that china fucks this up

You better bet someone started their agents with a prompt "Make a salesforce clone but with 100% uptime"

thought this link is better / cleaner: https://unwall.app/www.economist.com/science-and-technology/...

Yeah only somtimes it rewrites the whole files and thats the models/developer instructions fault. a good AGENTS.md is all you need, change my mind

still half as cheap as in EU

gpt-6-astra is a bitch, it constantly scope creeps itself with "yet another thing" to give it that darn polished lick. the results are eventually a little bit better but at what cost? let's do the math.

gpt-5.6-sol: 1x base gpt-6-astra 2.5x base in subscription

then gpt-6-astra tends to spawn subagents a lot, often with all kinds of models such as gpt-5.6, 5.3-codex etc., which is neat. it's a good coordinator but even more cost.

and then it tends to run _full test suites_ over an over again (each costs like 15 minutes) just to verify that _one test_ was fixed etc., and does so for as long as until the test is fixed, eventually accumulating 2 hours or so.

yesterday I assigned it a task to rebase my changs in a repo onto the latest upstream changes. while gpt-5.6-sol consistently took like an hour to do so end-to-end, astra ran for more than 6 hours and still wasn't done. it kept finding "one more thing" that was goldplating that I didn't ask for.

I consider myself an advanced industry representative in these things and would also borrow some time for detailed investigation. I'd happily contribute some valuable assets (tacos with cheese dip) into the matter, straight out of my personal drawers.

I use cachyos as my daily driver for a long time now. I play quite a few games. Only league of legends forces me to dual boot into windows every once in a week or so - fuck the anticheat cartel

Yeah that must be it. OpenAI doesn't want to disclose internal reasoning, that's why thats typically encrypted_content in OpenAI codex session ledgers etc.; leveraging responses API preserves reasoning server side all the way till a final answer is made; so that's very impressive and to me the score that matters.

French

Well, guess they are just trying to conform all the weird EU checkmarks

There are already a handful of repos doing essentially exactly this: `mlx-moe-offload`, `streamlx`, `mlx-moe`, `mlx-flash`, and `deepseek-v4-flash-mlx` - i.e. keep the resident parts of an MoE in unified memory and page/stream routed experts from SSD on Apple Silicon.

At this point I'd much rather see people collaborate on one of these implementations, benchmark against them, or upstream the useful bits into MLX/MLX-LM instead of producing yet another near-identical repo.

The local-LLM ecosystem really does not need every implementation idea rediscovered five times and wrapped in a new README. AI-assisted coding makes producing a new repo cheap; maintaining, benchmarking, and integrating one is the actually valuable part.

But I'm fine paying +50% more for +5% increased performance because it will pay off.

that model is 14MB large what do you expect. but I agree it's funny regardless

↲ is also two tokens instead of a simple \n lmao

So it's basically a container with a fancy name, innit?

how did this make it to the front page

75% quality... it gets worse the longer the context.

Really just a (not so) fancy CLI wrapper about a prompt and a skill: https://github.com/openai/codex-security/blob/f22d4a36f26d16...

I have a suspicion that they are just trained on puzzles by now

Funny timing. I've been building something similar in my spare time called Daydream. There’s a lot of overlap: local screen/audio capture, OCR and transcription, window and activity context, SQLite, and a searchable memory of what happened.

The main difference is the product direction. Screenpipe seems focused on continuously giving agents context through APIs, MCP, and skills. Daydream is more narrowly built around answering "what did I do today?" through a timeline you can inspect, replay, search, and turn into a daily digest.

I'm also treating deletion as part of the data model. If you cut a sensitive span, its frames, audio, OCR, transcripts, embeddings, and summaries should be deleted or invalidated too.

Mine is still early and Linux-first. I'm open-sourcing it in case anyone wants to contribute, poke around, or use it as a starting point. It’s built with Tauri, a Rust backend, React/TypeScript, SQLite, GStreamer, Whisper, OCR, and VLM processing.

I genuinely didn’t know you were building this when I started. Apparently personal memory capture is becoming a SaaS category too lol.

Code is here: https://github.com/snackbit/daydream