HN user
I agree the use of manual gearboxes in cars is decreasing. However, as of today, it is nowhere near a "lost art" and it won't be for a long time. There are enough manual gearbox cars in the world that, especially in developing nations, they will be around for a long time, even more so if agricultural and other industrial vehicles are considered.
The article was needlessly hyperbolic and narrowly focused on a single market.
Article is narrowly focused on the US market.
“Driving a manual vehicle is a lost art.”
Hardly. It may be decreasing overall but nowhere near a lost art.
“Yet manual-equipped cars still have a loyal following, making up nearly 29% of new-car sales in 2024 across Europe's big five markets.”[1]
[1] https://www.slashgear.com/2039491/why-manual-transmissions-a...
Which models have you found work best with OCaml?
I also went through quite a process to select a language to work with LLMs[1] before settling on OCaml.
I am not unhappy with the choice and find it works quite well, with relevant skills loaded, but I am always interested in others’ experience and understanding what they’ve discovered works well.
For the UK, there's a handy Wikipedia page with 57 defunct airports listed:
https://en.wikipedia.org/wiki/Category:Defunct_airports_in_E...
And also specifically former RAF stations (stations, airports and headquarters):
https://en.wikipedia.org/wiki/List_of_former_Royal_Air_Force...
Emmi is an Austrian lab specialising in physics AI applications.
Mistral isn’t specialising in French language LLMs either.
The point was that across different European countries and languages there are collaborations and M&A happening.
I received the same email, although couldn’t quite figure out which retiring model I was still using, as I thought I’d already transitioned to Mistral-Medium-3.5 for everything. Anyway, after receiving the email, my hope was that it meant they were also planning on releasing some new, improved models in the next months.
Mistral just acquired Emmi AI, an Austrian startup.
German and French speaking together at last.
I was a Mistral Le Chat Pro subscriber (the €20/month plan). Yesterday I hit my monthly limit. Switching to PAYG I burned through another €40 in one evening, working on the same project, with the same tasks.
I upgraded my plan last night to Mistral Le Chat Teams. This now costs me €60 per month for two users. Limits have been reset, but I have no idea now if my per seat limit is higher than the Pro plan, or if the limit is shared between the seats, it’s really not clear. I guess I will find out next month. The limits reset on the first of the month and I really hope I don’t hit them in the next seven days.
I use Mistral Vibe CLI and I’ve written and implemented a couple of new skills[1]. Caveman, based on an idea I found online somewhere, this skill removes all extraneous response text, including articles. Makes for some fun reading, but supposedly reduces output tokens significantly. Hash-anchors, this one is based on a concept from Dirac[2], reduces search failures and also includes multi-file dispatch. It will be hard to measure, but Vibe tells me these two should result in roughly a 40% reduction in token burn.
It's generally inconsistent. The first sentence is written, "A co-op is an economic system built on the simple idea that coordinating the economic activity..."
Co-op is correct here, but not in the title (Coop). Probably personal taste, but I'd also like to see hyphentation for "co-ordinating", "co-operate" and "co-ordinator" as well.
Then I noticed the em-dashes, so perhaps I'm reading the machine's work anyway.
I don't use the numpad characters, but I have tried Vibe, Goose (GUI and CLI), Dirac and the built in agent in Zed for vibe coding. I keep coming back to Mistral's Vibe. I actually find the ergonomics of it nicer than the others I've used so far. I really wanted to like Goose, and their GUI offering is OK for chat, but I thought their CLI was poor. Dirac was OK and I should try it again to be fair. Zed was just overkill and complex for what I needed. Vibe CLI seems to hit the sweet spot, although it's not perfect. The challenges I encounter are mostly down to API errors though and sometime bash tooling. I could configure it better for that, if I took the time (which I should).
You're thinking of 442 ;-)
Mistral’s Vibe CLI does support SKILL.md files.
Their Voxtral[1] speech models are really good.
Maybe they've fixed it since publishing, but the opening line now says:
Europe's leading artificial intelligence firm, France's Mistral AI, said on Tuesday it has acquired Linz-based Emmi
Of course ChatGPT is available to EU citizens, should they choose to use it. That’s very different to the Maltese government actively promoting use of ChatGPT.
Of course Maltese citizens can still choose not to use ChatGPT (until it becomes mandatory), but if the State supported education is bound to one particular tool, storing user data outside the EU’s jurisdiction, I think that’s something to discuss.
Malta is part of the EU. I am personally very surprised about this partnership, just in the context of data security, privacy and the GDPR. How is the privacy of these EU citizens protected when all their prompts and data is sent to OpenAI? How do these EU citizens submit a request for all their personal data to be deleted from OpenAI records, a right they have under the GDPR with a compliant data processor?
Also in this space is Moonbit[1] and Vera[2]. A performance comparison between them, across multiple factors, would be interesting because I have no idea of the pros and cons of each.
Does your goose talk to other geese?[1]
The initial criteria was strongly typed and functional first. Using an LLM for answers, of course, that returned me a list that looked like:
- Haskell
- OCaml
- F#
- Scala
- Gleam
- Purescript
- Grain
- Idris
Then I asked if there were any Schemes or Lisps that met the initial requirements, which added a bunch more options (Typed Racket, Typol, Elm, ReScript etc).
Then I asked about Julia specifically, as it's a language I'm already reasonably familiar with and knew that it's possible to write it with static annotations.
Next I started filtering the list based on additional criteria; didn't want to target a JS compilation target, performance, size of package ecosystem, tooling, community, learning curve (I do want to review and understand the output).
There were a bunch of follow-up questions over a few hours of prompting, reading and a couple of beers. All this resulted in the shortlist of OCaml, Typed Racket and Julia.
Julia pretty much remains in there, even though it doesn't really meet the strongly typed initial criteria, based on my familiarity, the ecosystem especially for AI/ML tasks and performance factors.
I know zero about OCaml and find the thought of learning it a bit daunting. Typed Racket seems more approachable anyway.
Thanks for that, I hadn't scrolled down far enough.
Just want to be sure I'm reading the results correctly... When I compare GPT-5.5 with Mistral Medium 3.5, I see in the tables:
a) Mistral beats GPT in Java and C++
b) It's close for Rust
c) GPT-5.5 easily wins for Go, Javascript, Python and Typescript
Model choice really does appear to be language dependent (assuming I'm reading the results correctly).
Just last night I was going down the rabbit hole of "what's the best programming language to use for vibe coding." I came to a short list of:
a) Typed Racket
b) OCaml
c) Julia
I would love to see those three added to your benchmarks. And Mistral Medium 3.5 added to the LLM list, please.
GeoJSON is not just for geographical features! Shapes of any kind work just as well.
QuPath[1], a tool for digital pathology whole slide image analysis, can export annotations in GeoJSON format (and import too I suppose).[2] This makes it really very easy to make annotations transportable between tooling.
[2] https://github.com/qupath/qupath-docs/blob/main/docs/advance...
Direct announcement on the Mistral website: https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-...
Nice to see some computer vision deep learning stuff here on HN, instead of more Gen AI.
However......
I didn't see anything on the website about whether this solution works for all skin types. Having worked extensively in this space (computer vision deep learning for histopathology, cytology and biology, not skin specific) there are always confounding factors and biases in these models when analysing samples from minority groups. How does this model deal with lighter or darker skin? What about age? Sex? Hairy or smooth? Was the training set varied enough to cover these scenarios? Maybe consumers don't care about this kind of information, but that's the first thing I went looking for (well, the second thing. The first thing I looked for was the model used; DINOv2 if I read correctly).
Buried in an FAQ it says, "Our clinical evidence package (Document ID: IVY-CLIN-001, March 2026) includes the full summary report, audit protocol, reviewer credentials, sampling methodology, and ICD-10 mapping. Available on request." I really think that's the kind of thing that should be available up front.
The website also says, "Built to international medical-device and information-security standards." Then there are little breakout boxes for ISO13485, ISO27001 and UKCA-Class 1. Are you truly an ISO13485 and ISO27001 certified company? Or just somewhat disingenuously implying that. If you are, then publish your audit certificates. (Tiny text says, "Standards listed refer to our implementation framework. Certifications are issued upon successful audit by relevant regulatory bodies.") Is your model a UKCA-Class 1 certified medical device? If so, why not IVDR and FDA 510k as well? I was pretty surprised to see the offering is from a UK/HK company, as the marketing is pretty squarely at a US audience in my view, including the pricing in dollars. What about HIPAA and how you handle PHI? No mention of SOC2, which is more widely recognised (asked for) in the US than ISO27001.
I appreciate you may have built a nice product, that works quite well in common use cases, but this is a tough and highly regulated space to play in. Good luck.
I am a Mistral Le Chat Pro subscriber. I specifically chose to test their offerings because they are European. I don't have the necessary local hardware to run really big models, therefore need to choose a cloud provider if I want LLM action.
I find the antics of Anthropic, OpenAI, Google, Microsoft distasteful and avoid their products where I can.
After testing Le Chat and Devstral-2 for a while, I felt their offering was good enough to stump up some cash for it. I appreciate that many of their models are open weights and Apache 2.0 licensed. In general, I've been happy enough with the service and quality.
Maybe others are better, but I have little reason to change right now. If curiosity gets the better of me, I'll be looking at Qwen, Kimi, GLM, Deepseek, other open weights models, before Anthropic and OpenAI.
I can not pay with a bank note suddenly?
In some places you cannot. I was in London post-COVID and there were a bunch of tourist things, like a riverboat on the Thames, where you could only pay with a card. Went to a craft cider bar out in the countryside and again, they didn’t accept cash. Personally, I think businesses should be forced to accept all legal tender, which means cash stays as a first class payment method, but that’s not how it is in many places.
On the other hand, in Austria there are many places that are cash only, especially small restaurants in the countryside or community sporting events with coffee bars.
Sorry, I can’t help with that, as I’ve never actually used Claude. Started with Mistral and stuck there (they still have a free API tier, but I ended up buying their Le Chat Pro service anyway, for the image generation).
These are already available. Mistral’s Vibe CLI[1] is open source. Tools like goose[2]are API agnostic.
I like it, but I haven't used Opus so it's hard to compare.
From the benchmarks I've seen, Mistral's offerings are right up there with other open source options, but not quite at the frontier with closed source models.