LeBit

It does look nice and easy.

Not sure I will give up on smolvm though.

I have scripts to launch an instance per task. Nono wraps my coding agent.

I provide the git clone for the specific task.

It works really well.

There is also GLM 5.3 Flash

What are you talking about?

DS4.1 Flash not really cheaper than frontier models???

It is insanely cheaper.

I have subscriptions to OpenAI and Claude but use DeepSeek 4.1 Flash for my coding agents.

It costs pennies and you got really great output.

The author is spot on.

That is always amazing to me.

There is no way I can beat even local models at generating complex Python scripts fast.

hn is filled with uber geniuses.

Don’t want to be an asshole but the pattern I have observed is that people who were not qualified / competent to code before November of 2022 have this mentality.

They want to make you feel dumb and useless with your software development knowledge.

smolvm uses only the required memory and releases what isn’t needed anymore

Smolvm

You don’t mind letting a company get 100% access to your personal data?

Unless you use a zero data retention inferencer?

After looking at so many options, that is also my take.

These should be decoupled.

Maybe I need nono in one context and smolvm in another or both.

I would not want to trust the harness to self policy.

Yup.

I see posts about OpenAI and Anthropic latest and don’t even care looking at what they do better. I just read the comments here.

I use DS4.1 Flash and GLM 5.3 Flash, pay peanuts per day and get more than acceptable results.

Let’s pray Chinese models are not banned.

This is close to saying LLM users are psychopaths.

> Yeah, I'm in Montreal right now and have been living there for quite a while.

shocked Pikachu

Tes interventions montrent que tu es un raciste intolérant. Je n’ai plus de temps à perdre avec toi.

Not really. They are located in Europe. Share a border with France. Still have a lot of laws to protect French.

Your comment is dripping with racism.

Did you know there are tensions in Switzerland because there is a will to stop kids from learning French while they attend primary school? https://www.edi.admin.ch/fr/newnsb/CbsqmmLpkhDjQgEILMp_P

Did you know Switzerland is located in Europe and shares a border with France? Did you know people in France speak French?

Do you know where Quebec is located?

I’ve seen your kind before.

128G of RAM? I envy you.

You will be able to run some pretty impressive models locally. Good for you.

"Extremely illogical". Okay buddy.

What other place do you know of whose context is the same as Quebec?

Je t’assure que ce besoin qu’on certains français de montrer leur capacité de parler l’anglais, de nommer leurs entreprises avec des noms à connotation anglophone, etc. est considéré très gênant ici du Québec.

J’imagine que c’est de plus en plus facile de le faire avec l’aide de l’IA!

I’ll look into this. But there is no universal solution. I find this to be quite limiting. I understand this is on purpose. If you want to automate, you need to pay more to have an api key.

It’s a good thing we have access to those very cheap and useful Chinese models.

I use the following models:

  - Qwen3.5 9B
  - Qwen3.6 35B A3B
  - Qwen3.8 27B
  - Gemma 4 26B A4B
  - Gemma 31B
  - Muse Glimmer 30B

I have 48G.

The MacMini is used solely for inference. llama.cpp + llama-swap.

People are saying local models are crap and serves no purposes.

I use them to help me spell check, write emails, write text messages, write JIRA tickets, write PR comments, etc.

I also use them in coding agents to complete different tasks.

I find them quite useful!

I use local LLMs on my Mac Mini.

Otherwise DeepSeek Flash 4.1 is dirt cheap (other "Flash" models are not that expensive either). I pay (very few dollars) out of my own pocket.

There are many things where having an API Key is necessary.

Maybe I’ve missed the boat though: is there now a method to use an api key to access a subscription?

Didn’t Israel use AI to target Hamas agents?

You would think, following the absolute horror that was, that any civilized nation would not use such a tool in such capacity.

Or rather, it showed it was working wonderfully for cases where war crimes is a goal.

Check out nono.

They already exists. You haven’t been invited? Hmmmmm

It’s not that bad: https://news.ycombinator.com/bestcomments

Why they instead of a microVM?

What do you mean by "It doesn’t provide isolation" ? How so ?