Mashimo

I think they all support branching if you use a agent like pi.

Seems like in agentic work flow the qween flash and Deepseek flash models are quite good.

Fits with another comment from yesterday on here who said the flash models are just better at tool calling.

Planning with gpt55 and implementation with a flash model could be bang for the buck route.

At least v3 did not when run selfhosted.

Why are you asking?

And yet, look at the French one. Can't compete with one year old open weight models even though they just released a new model this week.

Probably best to look up the local prices yourself? We don't know where you live.

There are fans that are cheaper that come close to noctua, but noctua are one of the best fans you can buy.

> It's not the same operations used to write code

Is it not? It's html and javascript. And not even attempting to draw details that other models do.

When I try other html / js prompts it also lacks behind china models from over half a year ago. I mean worse then GLM 4.7.

I can't speak for what you consider sucking, but there is a significant difference between Mistral and Kimi or Gemini. I find the others to be usable for my needs.

It's so bad I don't want to spend the 18 EUR just to test it for a month. It can't even create an SVG of the facebook logo. There should be plenty of examples of that around.

Gemini fast could do that in under 5 seconds.

I also did some SVG tests, it's really bad.

https://chat.mistral.ai/chat/897fbe7d-b1ae-4109-9b29-f3ccc4f...

Compared to all other hosted LLMs that I have tested, Mistral seems to be the only one with rather strict CSP headers. When you ask them to create a website with some javascript library it will not preview, even though le chat offers canvas mode.

Sometimes when a new release comes around from any provider I just want to test it a bit on the web. without paying and using an agent harness.

Why are they like this ;_;

Edit: Christ on a bike it's bad at drawing SVGs https://chat.mistral.ai/chat/23214adb-5530-4af9-bb47-90f5219...

> I think you’ve got it backwards. GitHub is by far the market leader for hosted repositories

Maybe, but I never heard about any company using github for internal projects in my real life. For me it was always to go to for open source projects.

Then again it's not a topic that often comes up in my developer circles.

> I guess it’s all the other non-git stuff

Yes, you want to run automated builds, unit test, end to end test, UI tests, make it easy for testers to deploy specific versions / tags to internal server. Also kick off builds for iOS on mac computers. We use Teamcity for that.

Tracking of issues, feature and epics. Maybe also knowledge base / wiki. We use Jira.

And pull requests. Bitbucket.

Demo of Gerrit here: https://gerrithub.io/q/status:open+-is:wip

Jira / Bitbucket / Teamcity.

Might be pricy though.

> The UI and concepts are familiar.

I guess, but it's not like you can't learn how to create a pullrequest on bitbucket or how to create an issue on jira as well within a work day?

That seems like the smallest thing when switching to a new company.

> The friction for adopting features like Actions is relatively low.

Yeah, I know almost nothing about the CI integration and actions when it comes to Github. Will look into it. Thank you.

Oh FOSS projects I totally understand. It's where I go to too.

But closed source companies surly don't need to establish a community?

What issues do you have with bitbucket?

Honest question, why are companies interested in hosting on github?

As a private person I use it too as a free hoster, but from work I mainly know self hosted instances of jenkins and TeamCity.

> A report finds a third of newborns use devices for more than three hours, despite government advice that under-twos have no screen time at all

Disgusting :(

That said, I can't read the article, paywalled. Anyone have a working link?

> https://code.overheid.nl/RegelRecht/regelrecht

> Machine-readable Dutch law execution. regelrecht takes legal texts, encodes them as structured YAML, and runs them as deterministic decision logic. The engine takes a regulation and a set of inputs, evaluates the decision logic, and returns a result with a full explanation trail

Can someone explain this to me? Not the technical aspect, but rather a user story or use case, maybe with example. I can't really wrap my head around it. Thanks in advanced.

I have 70+ year old relatives who use Signal. It works quite nice. Similar to whatsapp.

He is saying the same as you :)

I think it's an ok counter argument. You can't have "AI should do the users bidding" and "implement a sensible detection and reporting pipeline."

I mean that is what anthropic tried here.

You will not get banned if you use the API. AFAIK you can't use the subscription with other harnesses. That is how I understood it.

Wow, that is .. quite some thing.

Here in Denmark we have MitID, which you kind of use for everything. Taxes, every private banking login (you switch banks, keep the login), signing documents, all kinds of healthcare websites and apps, student loans and what not.

Not the one you asked, but the github has a single example: https://www.youtube.com/watch?v=VH9g66e42XA

You could also just use it as a voice interface, not touching the keyboard and just say "Compute, add reverb to my chords" "Computer, switch 808 drums out with 909 drums but keep the pattern and effects"

But how do you authenticate when filling your taxes?

> They sent my refund to some identity thief.

How does that currently work?

In DK they just send it to your "nem konto", the same bank account that also gets your wages. More or less a sym link, so even if you move bank it will follow. Makes life easy.

Probably like the children game "Broken Telephone" or "Gossip" where after a long chain of word of mouth input does no longer match output. Sprinkle communication between different companies on top of that.

Thirty Seconds to Mars -> "The Mars band" -> ?? -> Bruno Mars