Hacker Newsnew | past | comments | ask | show | jobs | submit | baq's commentslogin

MCP vs CLI is the same discussion as between a GUI app and a web app: it's all about the distribution. There is approximately no difference in functionality except whether you're hitting a dedicated service or running a local tool which connects to a dedicated service.

With saas is turned out that distribution to a browser solves a pretty major pain point and I expect MCPs to be treated the same. Can you trivially replace an MCP server with a CLI tool which accepts a token? Yes - but why do that to yourself when you can hit the endpoint directly?


They can get what, 1B euros? 10B when everyone loses their mind? This doesn’t buy nearly enough compute nowadays.

Meanwhile, Anthropic and OpenAI have investors practically begging them to let them buy this much equity at mind-bogging valuations.


The chinese labs manage to do it. Mistral should have enough money.

The EU has intentional structural hurdles to pouring money into a predetermined single company. Both hurdles meant to fight corruption and nepotism, and hurdles meant to ensure fairness between the member states. After all, money to Mistral is money to France too, and you don't want countries to abuse such mechanisms

It's not impossible, but China is just much better set up for the nessesary level of government support


China is a way more corrupt country but this might be a benefit of less rules.

China has cheap coal powered electricity and leaders that make things happen. Europe has beaureaucrats that only love talking, high taxes and expensive energy.

[flagged]


You think they're intentionally being bad because they can't manage to pump $65B into a startup on a whim...?

You think well over a year after making grandiose on the record claims is “on a whim”?

What claims are you talking about?

I've never heard or read anything about the EU planning on investing money in Mistral. They're a private company. They're French. It honestly sounds kind of absurd.


[flagged]


No, we want you to backup your claims and provide sources or stop adding pointless low effort anti-EU noise to the conversation. It's frustrating, any time there's any kind of discussion about anything European on HN it gets flooded with shallow, low effort "EU-bad" posts like your contributions here.

If you're going to make that claim at least put some effort in.


This is a mostly American forum and some people want to piss on the EU to elevate themselves. Europeans do the same to the US but about politics, health care, work life balance, and quality of life. You know, the stuff that matters :D

From what I can see you put in zero effort in a response and you expect me to put in more effort?

I already checked for one variation of a google search like I said.

Can you show some proof you did anything at all?


Or maybe they’re just poor.

Low key amazing tech, kinda like clickhouse - nobody is bragging it’s running their business

agreed, the next price increase from frontier labs (and the inevitable limits decrease in subscription tiers) will have people thinking real hard about their model providers and that's when mistral should be ready. however, given their recent performance, I realistically don't have my hopes high up.

DeepSeek is both cheaper and better than Mistral.

Not in many tasks. I use deepseek as a fallback in https://phrasing.app and it’s always very apparent when it happen (due to mistakes/clear performance drop off)

Interesting - which models specifically? I'd be interested in using mistral over deepseek if it was competitive (guess I need to go benchmark)

I use small, large, an medium-3.5 depending on the task

Because they distill

I feel like there's an implication here that distillation is a problem but I don't understand what you mean. I thought distillation was generating text from a model and then training another model on it. Is the something unethical in that? You're paying the API costs to generate the tokens, right?

Or I guess more to the point: is this something frontier labs have said is (or tried to paint at any rate) problematic? This feels like an "out of the loop" situation because I've only ever heard "distillation" with a positive connotation before.


Whether it's a 'problem' or not is viewpoint-dependent but it's against the OpenAI ToU:

> You may not use our Services for any illegal, harmful, or abusive activity. For example, you may not:

> [...]

> * Use Output to develop models that compete with OpenAI.

Source: https://openai.com/policies/row-terms-of-use/

(I'm also curious whether they consider developing a competing model to be illegal, or harmful, or abusive...?)


> it's against the OpenAI ToU

Given that OpenAI doesn't care about training on copyrighted data, why is suddenly their ToU something anyone should care about?


That OpenAI was in the wrong when they ignored everyone copyright, does not make it right to ignore their ToU. If a one wants IP and rule of law (incl contracts) to be respected, one should not violate others rights when it is convenient.

On a more risk-strategy level there is the size of their legal team, general endowment, and supplier and political connections to consider.

Everyone is free to ignore their ToU, but I can understand why a company would avoid it...


> If a one wants IP and rule of law (incl contracts) to be respected, one should not violate others rights when it is convenient.

Yes that's what should be said to OpenAI. Now they should not cry about their T&Cs not being respected when they never cared about others' copyrights.


Feels like this should be some kind of anti-competitive violation even if it's not actually. Probably moot under this admin but still.

It's like saying you can't use windows to develop an OS, or drive a Ford on the way to your job at Hyundai.


In heavily-unionized areas, you'd best not drive that Hyundai to your job at the Ford plant.

it doesn't matter the reason. This is a race and nobody will care or remember how the winners got there.

Mistral looks like it's fading away to irrelevance unless they can play alongside the similar sized models, or have some unique advantage other than being in Europe, for Europe. I was really excited for them back when they were startup that had the biggest European venture round ever. This space will have a few winners, and many losers. Google, plus either Anthropic or OpenAI most likely. Big models will see breakthroughs in inference performance/cost fall precipitously and small models will only exist on devices (Pixels and iPhones, cars, watches, bluetooth speakers, etc)


It’s not that I don’t agree with you, I am just pointing out why it’s hard to catch up to scaling laws given the European economic (capital) and political (US would be upset if they found out Europeans distill) constraints. China is only bound by economic constraints.

With the insight in your comment and this bit from the above one:

> This is a race and nobody will care or remember how the winners got there.

It seems like the EU should have paid China for the distillation datasets, esp. since Mistral isn’t even a governmental org.


> This is a race and nobody will care or remember how the winners got there.

For consumer AI, yes. For coding assistants, probably.

For specific application "business" AI like the things Airbus announced the other day? Not at all. What matters for an Airbus using Mistral to build compliance documentation based on AI generated physics simulations is the enterprise relationship, reliability, compliance, forward deployed engineers helping with the fine tuning, quality, predictability, support. A Chinese lab having a better at benchmarks model that is cheaper is just irrelevant for that.

And IMO, the real money in AI is this type of "business AI" deployment. Developer tooling tends to converge on becoming commoditised. Once you're a core supplier for a big bank and embedded in their processes, you're there untill you screw up with the pricing (like Broadcom), and even then.


Why doesn't Mistral distill?

Good question, given that American companies basically threw copyright law into the trash, I think they should.

American companies can't sur Chinese ones, but they can do it with European ones.

So then the European ones should join with European copyright holders to sue OpenAI/Anthropic and watch them trying to BS their way around what they train on.

Well if they did wouldn't be able to feel superior to Americans about that particular thing. Perish the thought!

It’s really a pity, why can’t they feel superior while breaking ToS and copyrights just like Americans can feel superior over Deepseek while breaking ToS and copyrights?

I suppose losing with dignity is a consolation.

Also, new Medium 3.5 is far more expensive than previous Mistral models, and much more expensive than e.g. Deepseek

I tried it out on some dev tasks with their Mistral Vibe subscription, and the performance was pretty okay (okay, not great), both in regards to development and speed. Worse than Anthropic's models I'm used to but at 20 EUR per month it wasn't a bad deal - except that the 200k context size would more or less be a deal breaker in many cases.

Where do you sign up for that subscription?

I wanted to try out Mistral, but I fail to find anything like that even after creating an account


The other comment already mentioned that you get their subscription: https://mistral.ai/pricing/ they do say that you can try out their coding agent for free, but personally the Pro tier is pretty affordable too to try out for a month.

Then you can install their coding harness, I personally used the Python + uv option: https://mistral.ai/products/vibe/code/ if you don't have uv yet, you might have to install it too: https://docs.astral.sh/uv/ though I already use it for other projects. Oh and if on Windows, you probably want to do all of the installation inside of WSL, just so that file paths are the *nix variety, I've had issues otherwise with pretty much every coding harness, like OpenCode as well (across multiple models).

After that, you need an API key for your subscription, you can generate and copy it here: https://console.mistral.ai/codestral/cli that's also where you see the quota, though it seems to NOT refresh instantly, but more or less a few times a day.

Either way, happy coding!


Maybe on their pricing page?

https://mistral.ai/pricing/


Everything is more expensive than deepseek. They aren't frontier in intelligence but they are the frontier in cost per intelligence

I'm yet to see a linux distro with memory configured correctly out of the box. (I haven't looked too hard, but the defaults are abysmal.)

Still can't help the fact memory management on macOS is vastly better with its use of pages compression and unlimited swap.

Interesting. I'm now working on some admin scripts and will add this to the list.

The title should read 'it's hard to justify buying any other laptop than the Neo in the sub $1000 space'. It's an absolute unit of a computer; the only more revolutionary box would be the M1 Air (or the original Air. maybe. my vote is on the M1.)

The original Air was not good.

I think you mean the second gen Air (SSD-only, c2010), which was an incredible combination of price, performance, and usability.


With the software supply chain running amok recently having anything connected feels like playing Russian roulette and I say this as somebody who is running home assistant for years. I’m particularly paranoid about connecting my ev (non-vw) to it now, feels like a serious footgun today, would’ve been convenient three months ago, true.

Yeah exactly. Blowing up the rocket is the easy part. Reliably blowing up rockets on a high cadence is hard.

On the scale of bad 1-10 where 10 is the absolutely worst case this is a 12 easily.

(Elon’s strategy of blowing up smaller versions of their rockets more or less deliberately doesn’t sound so insane in the light of this.)


I'd say on a scale of bad 1-10, 9 and 10 are reserved for incidents that cause loss of human life. YMMV.

Loss of human life in a static fire is criminal. Why would anyone be that close?

There was no loss of life in this static fire failure.


I can think of a few reasons:

- Test commences prematurely when people are still around

- Test is aborted partway through but then spontaneously resumes when people have started coming back

- Error in design or failure of hold-down structure turns static fire into dynamic fire, moving fire to where people are

These are unlikely, of course, but they are the things we have to seriously think about and try to design out of the system in order to create safe systems.


No one should ever be that close, but it's a worst case scenario within the realm of possibility (people do get themselves into danger sometimes, for example by wandering onto a railroad track when there's a train approaching). I don't think it's unreasonable to reserve the 10 on the 1-10 scale for 'loss of human life'.

  The train driver saw a man on the track ahead holding a cell phone to one ear and cupping his hand to the other ear to block the noise.
https://darwinawards.com/darwin/darwin2002-24.html

> Loss of human life in a static fire is criminal.

True. And yet it is not without precedent.

Scaled Composites had an explosion while performing a cold flow test of SpaceShipTwo’s engine which killed 3. https://www.latimes.com/archives/la-xpm-2007-jul-27-me-explo...


I mean, there was that one static fire recently where the rocket broken loose and started flying. This was not for from a populated area. Ok, maybe that was pretty criminally negligent.

The Plainly Difficult channel on YouTube reserves 1 and 2 for incidents that don't cause loss of human life.

yeah if you want to put it in the best light in terms of 9/11's this is zero 9/11's of casualties. Not how I'd judge it.

SpaceX had a very similar failure during a static fire test in 2016 that destroyed the rocket, payload, and a few key parts of SLC-40 that took them over a year to repair and return to service (September 2016 -> December 2017). The concrete flume trenches were literally melted.

https://spaceflightnow.com/2016/09/01/spacex-rocket-and-isra...

That was a full size rocket on a real mission with the $200M payload on board during the static fire, which is ostensibly worse. The payload was not integrated yet in Blue Origin’s case.


None of this matters in the product: it either is capable of agentic loop workflows or it isn’t. A 10% improvement in probability of single task success makes or breaks the use case.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: