Hacker News Reader

toshon Dots: Always-on agents

> excluding the European Economic Area, Switzerland, and the UK

https://help.openai.com/en/articles/20001530-getting-started...

toshon OpenAI: Tomorrow we are re-opening the Pro $200 subscription

> Over time, we see prices go low enough that it makes sense for most to buy usage as needed

toshon How to keep enjoying programming in a world of LLMs

Pick something you always wanted to learn more about and use LLMs to help you

toshon We must pace the frontier

I also have a proposal

Anthropic releases models as open weights + more information about how they do training and alignment

toshon The Navier–Stokes Millennium Prize Problem

> My two favourite hypothetical questions regarding this used to be:

> If I'm running Codex and one of my API keys accidentally gets consumed in the context, what are the chances that someone else might ask for an API key in the future and get mine back? (I asked someone at OpenAI once and they called this the "regurgitation" problem and assured me that they take great pains to prevent that... but wouldn't describe how.)

> If I brainstorm with ChatGPT about potential new directions for my company, what's the chance that information might be exposed to a competitor in six months' time who asks "what might company X plan to do next"?

> My new preferred hypothetical for this is:

> If I use ChatGPT to help me partially solve a Millennium Prize problem, what are the chances that my work will influence training such that a later model helps someone else solve it first?

toshon GPT-6 Astra in code review: Gains, privacy, and cost

fwiw I found Astra to be faster than Sol w/ both on medium reasoning for simple agentic coding

difficult to compare though because for more open ended, complex tasks Sol might miss something that Astra notices and then Sol might yield a cheaper but worse outcome

toshon GPT-6 Astra in code review: Gains, privacy, and cost

I ran a few toy benches comparing Astra with Sol

and found Astra ~30% faster and at similar cost to Sol for the same outcome

https://x.com/__tosh/status/2096201900555170032

the token efficiency helps Astra even though sticker price is 2.5x that of Sol

toshon GPT-6 Astra

$10 per million input tokens and $50 per million output tokens

sol is $4 / $20

toshon OpenAI begins rolling out GPT-6 Astra

not released yet

toshon Claude Fable 5.1 and Claude Mythos 5.1

> The watermark doesn't change the meaning, quality, or readability of the output

how?

toshon AI Agent Has Root

you can run the agent in a different environment

toshon Migrating to HTTPX2

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

toshon Migrating to HTTPX2

nb: operating system TLS trust store is used now (instead of certifi)

toshon Small Models Have Arrived

I think we'll see more of this soon

replit is already leading the way with free luna usage

toshon Harness Engineering

the one by https://earendil.com/posts/what-is-a-harness/ is also a good read

and if you want to see what a simple harness implementation looks like you can just dive into the code and ask your favorite agent about it

https://github.com/smol-env/smol

https://github.com/earendil-works/pi

toshon Harness Engineering

it lists 'Garbage Collection' as one of the 3 components of a harness

I don't know any harness that solves 'Garbage Collection' in the way described here

(most harnesses accelerate context pollution and code base drift via instructions they embed into system prompts, tool descriptions and skills)

toshon Choose Boring Technology

oh good catch! I still had it in my open tabs

toshon Tim Curry has died

underrated movie and Tim Curry role:

Herkermer Homolka in Congo (1995)

https://www.imdb.com/title/tt0112715/

RIP

toshon Qwen3.8-Flash-Next

this is a new architecture (foreshadowing qwen 4)

> trained at just 1/9 the cost of Qwen3.7-Plus, while outperforming it across the board

https://x.com/Alibaba_Qwen/status/2092591393424515114

toshon Qwen3.8-Flash-Next

this is a new architecture (foreshadowing qwen 4)

> trained at just 1/9 the cost of Qwen3.7-Plus, while outperforming it across the board

https://x.com/Alibaba_Qwen/status/2092591393424515114

toshon AWS Acquires DuckLabs

did not expect that

I hope the future for DuckDB is still bright

toshon Z.ai confirms Ox Alpha is a new GLM-series model and will release its weights

my guess is this is a small model punching way above its weight

on toy benches it made quite a few mistakes but was able to fix all of them on its own

(meaning more tokens, more turns, more tool calls — but same outcome as gpt 5.6 sol)

toshon Omacom Foundation funding hits $10M

there are plenty of devs that are on macOS because they prefer a 'just works' + good GUI combination

it's good for linux if there is more activity on that front

there is a reason why ubuntu and fedora became so popular

toshon Fast and Hard Code

you can also situate the agent within your codebase and have it only write things you review (or have it not write/change the code at all)

toshon Fast and Hard Code

what i'm seeing is agentic coding helps be(come) more ambitious

build your own framework, database, operating system, game engine etc

things that used to be infeasible (too hard, too big, …)

toshon ElevenLabs, TwelveLabs, ThirteenLabs

https://www.youtube.com/watch?v=JB2di69FmhE

from There's Something About Mary

toshon Activation Energy is a good model for a lot of things

when you want something to happen (more):

- increase energy for activation (motivation, habit, …)

- make obstacle lower (break it down, find an onramp, …)

sometimes one is easier than the other

other times working on both is what makes it happen

toshon Berd

nb: this is by block (https://block.xyz)

toshon Berd

green is not a creative color

toshon Building Without Predicting

also a great approach for writing software