More

LeBit · 2026-09-21T00:39:51

How does it compare to kagent (https://kagent.dev/)?

LeBit · 2026-09-21T00:35:08

smolvm integrates with k8s (https://smolmachines.com/docs/guides/kubernetes-in-a-microvm...) and can mount s3 buckets (https://smolmachines.com/docs/local/machine-lifecycle-cli-re...).

I wouldn’t dismiss smolvm so fast. It brings together many ideas that make the whole very interesting.

srcreigh · 2026-09-21T01:05:46

It doesn’t provide isolation. It is not even part of the conversation.

LeBit · 2026-09-21T00:20:14

For microVM, smolvm is quite impressive.

For further isolation, I like to use nono inside a smolvm instance.

LeBit · 2026-09-18T21:33:42

I always thought it was bold for people to post anything using their real identity.

You never know if a stance you have now will be frowned upon in 5 or 10 years from now.

I’ve heard you can’t pass customs with a phone without your social network accounts.

mywittyname · 2026-09-18T21:52:31

You don't need to post under your real name to be IDed.

It's trivial to acquire a list of real people tied to email addresses, phones, social media. And that's what's out there for "normal" businesses.

Who knows what kind of tooling is available to people who serve on the boards of tech giants?

derwiki · 2026-09-18T21:37:01

Maybe not all customs are created equally, but flying between Toronto and the states I have never had to show my phone let alone my social accounts. Also never seen that going up through customs.

kmoser · 2026-09-18T22:29:54

Sometimes people get searched randomly, and occasionally it correlates to their skin color.

LeBit · 2026-09-18T21:26:20

Land of the … laugh … free!

LeBit · 2026-09-18T12:05:52

What is Astra $/task?

Even if OpenAI end up using 1 token for per task, if the token costs 1M$ , some people will find it expensive.

nater5000 · 2026-09-18T17:07:08

https://artificialanalysis.ai/#intelligence-comparison-tabs

Astra on xhigh has a cost per task of $2.31 with an intelligence index of 53. Qwen3.8 Max has a cost per task of $5.41 with an intelligence index of 45. Pricing for GPT-6 Astra (xhigh) is $10.00 per 1M input tokens and $50.00 per 1M output tokens. Pricing for Qwen3.8 Max (0902) is $2.00 per 1M input tokens and $6.00 per 1M output tokens.

Obviously this is just one measure of all of this (and Qwen 3.8 Omni Flash isn't yet available), but I think this illustrates the point well. These relative task costs are pretty consistent across different analysts. Cost per token is arguably a useless measure at this point in most circumstances.

LeBit · 2026-09-16T12:40:01

The USA would take them by force because there is no way anyone in the world wants to have anything to do with the USA.

But taking someone by force in on brand for Trump, so it might happen.

LeBit · 2026-09-16T12:38:29

Je suis tellement heureux de constater que Carney bouge rapidement.

Au début des annonces des tarifs, plusieurs experts disaient que de se découpler des américains étaient impossible à court terme. Nos économies sont trop entrelacées pour changer de cap rapidement.

Je ne sais pas exactement ce que Carney a en tête , mais j’espère que ça va accélérer le découplage.

52-6F-62 · 2026-09-16T17:30:16

Il a écrit un livre dans lequel il aborde ce qu'il a dans la tête.

C'est une bonne interview:

https://www.youtube.com/watch?v=1g0nLGZIzhc (Anglais, désolé)

Mais les cas les plus extrêmes en Amérique sont relativement récents...

LeBit · 2026-09-16T12:34:06

Trump was elected twice.

Until the USA fixes their broken political system, it would be unreasonable to seek to create major bonds with them.

They will remain a partner, but they should be a minority partner.

LeBit · 2026-09-16T12:31:18

No. The US has shown it is completely unreliable.

Even if Democrats were to be back in power and devote their 4 years to restore things and apologize, that wouldn’t be enough.

Americans have shown they can vote TWICE for a fascist regime that is willing and eager to generate chaos.

Their political system is foobar.