> Streaming speech recognition running natively and in the browser. A pure Rust implementation of Mistral's Voxtral Mini 4B Realtime model using the Burn ML framework.
> The Q4 GGUF quantized path (2.5 GB) runs entirely client-side in a browser tab via WASM + WebGPU. Try it live.
Excluding names (Mistral's Voxtral Mini 4B Realtime), you have 1 pretty normal sentence introducing what this is (Streaming speech recognition running natively and in the browser) and the rest is technical details.
It's like complaining that a car description Would contain engine size and output in the third sentence.
> Literal scum of the earth engaging in coordinated harassment campaigns to get people they hate kill themselves, and celebrating their "success"
Sure, the problem with Kiwi Farms is that people "disprove of them", not what they disprove of. KF were even blocked by CloudFlare, who have a very strong neutrality policy, that's how toxic, hateful and illegal they were.
Sometimes they hallucinate them, or if they exist, sources include blatant nonsense (like state owned propaganda, such as RT) / don't support the claims made by the output.
wtf are you conversing with LLMs that you regularly are running into "state owned propaganda" in the references? my "blatant nonsense" detector is going off...
My favorite is when it cites 5 sources, and 1 of them is a real source, and then the other 4 are short form junk that point to the first one as the source. So basically its just picked one article and summarized it for you and not picked any info from any other places. Oh and bonus points when I type the exact same prompt into a search engine, and that 1 source is the top search result anyways.
There is nothing wrong with Wikipedia, at least in the main languages. It's crowdsourced and has citations (and where there aren't "citation needed" help identify that).
It gives you superficial, in depth and factual information, with links to sources for more details if needed.
There's nothing at all wrong with Wikipedia but it needs sources to cite since it doesn't allow original research and the World Factbook is an important one.
Looking at the underlying study, this isn’t evidence of bias. It’s evidence of correlation between Republicans and negative sentiment.
If you look at the sentiment for public figures given, the bottom one is, for example, Brett Kavanaugh. Well, he was credibly accused of sexual assault during his confirmation hearings, which was a huge deal at the time. Someone with that on their record will probably be read as negative, but, I mean, not the editors’ fault!
> His policies resulted in the deaths of tens of millions of people in China during his reign, mainly due to starvation, but also through persecution, prison labour in laogai, and mass executions
What's "kid gloves" about that?
Let's contrast with the the farthest thing from a leftwing dictator we can find, the quintessential rightwing one, Adolf Hitler. Here's the intro to his Wikipedia page:
> Adolf Hitler[a] (20 April 1889 – 30 April 1945) was an Austrian-born German politician who was the dictator of Germany during the Nazi era, which lasted from 1933 until his suicide in 1945. He rose to power as the leader of the Nazi Party,[b] becoming the chancellor of Germany in 1933 and then taking the title of Führer und Reichskanzler in 1934.[c] Germany's invasion of Poland on 1 September 1939 under his leadership marked the outbreak of the Second World War. Throughout the ensuing conflict, Hitler was closely involved in the direction of German military operations as well as the perpetration of the Holocaust, the genocide of about six million Jews and millions of other victims.
When it comes to politics and studies... We all should know by now to research those sources too, right?
"The Manhattan Institute for Policy Research is an American 501(c)(3) nonprofit[5] conservative think tank."
It is a report generated by this conservative organization (that presumably gets donations from many other conservatives). Is there a chance that the report itself is suspect?
"Bias" here is just sentiment analysis. The report (from a conservative think-tank) is not about factual errors. Plus, the effect they find shows only for US politics, where there is really not much of a "left".
> Findings show that Wikipedia entries are more likely to attach negative sentiment to terms representative of right-leaning political orientation than to their left-leaning counterparts
Is that a bias or just reality?
Right leaning politicians in the US include people paying underage girls for sex, people screaming about "Jewish Space lasers", people obviously stealing money in plain sight with crypto pumps and dumps, people running away from responsibility, people getting caught engaging in sexual acts in public, and on and on and on. Their left-wing equivalents are... extremely mild by comparison. What, some run of the mill corruption and sexual comments that resulted in resignations?
If go past "right wing is associated with more negative things", and look into what those negative things are, you'd realise it's just reality. Just because there are two parties and two categories of political leanings doesn't mean they are somehow equal.
It is bias if your "factual reality" over-exaggerates the "facts" for the "bad guys" and under-exaggerates / completely neglects to report on the "good guys"
But to put it with John Steward, what if reality itself has a left leaning bias?
What if left leaning people have empirically broader empathy [0] which could imply that right leaning people have in tendency worse personalities. I guess you would attest yet another biased article here.
There are also genuinely good guys and bad guys. Reality, itself, has a bias. To think that ideology doesn't correlate at all with how moral you might act is, frankly, stupid. Not all positions are created equal.
Obviously you've never seen French software, so why pretend like you have?
Counterpoints: Deezer, Doctolib, Back Market, Tidal, Adopte, Mistral, Dassault Systemes (the company behind the two main CAD softwares out there), Thales, Qonto, Kyutai, Mirakl, BeReal, Klaxoon, ABTasty, etc etc. We can do this all day.
Oh, and there are ton of official government open source projects.
And no, "but they're not as big as a FAAG" does not mean that the software isn't good or innovative.
They are saying that France is the only country that builds its own defence stuff only in country (unlike Germany that would buy American jets, for instance).
And it's true for most things that matter. The only exception I can think of is manufactured by cooperative companies that are part owned by France and have facilities in France but that specific thing might be made somewhere else (e.g. Airbus A400Ms for which the final assembly is in Seville, but Airbus is jointly owned by France, Germany, Spain), and small arms (HK, German made). Small arms don't really matter in the grand scheme of things due to the low barrier of entry both for acquiring off the shelf and setting up domestic manufacturing. Good luck doing that with fighter jets or cruise missiles.
airbus is owned by germany and france to each 10% and spain 4%.
Things get interesting if you count all the owners from the states together they have ≈ 12%.
But it is controlled by Germany france spain.
> Autoscaling is a solution to the self-inflicted problem of insanely-high cloud prices, which cloud providers love because implementing it requires more reliance on proprietary vendor-specific APIs. The actual solution is a handful of modern bare-metal servers at strategic locations which allow you to cover your worst-case expected load while being cheaper than the lowest expected load on a cloud
And how do you predict with certainty your "highest expected load"? And if you're in a space like ecommerce, where you have 1 week out of the year with x10 or x50 the load, I doubt it would actually be cheaper than using autoscaling. Especially today, with the costs of memory and storage. Not to mention that whenever you hit your load maximum, you have a few months of lead time to get extra capacity.
And FYI, "proprietary vendor-specific APIs" sounds very scary, but if you think about it for a few seconds, those APIs end at configuring an autoscaling group which is mostly about your min/max, and scaling rules. Yeah, it's proprietary, but it's 3-4 parameters to configure based on what you need, and from then little if any adjustment is needed. And you can take the same logic and port it to any other cloud provider within ~10 mins at most.
> And how do you predict with certainty your "highest expected load"? And if you're in a space like ecommerce, where you have 1 week out of the year with x10 or x50 the load, I doubt it would actually be cheaper than using autoscaling.
If you know when that week is, where's the problem in spinning up extra capacity just for that period, a week in advance?
Retailers, whether online or not, know with a very high confidence what their expected load is going to be a week from now.
Doesn't mean, it significantly complicates state. Because the place that stores it (say, your bare metal machine with your database) is now the bottleneck.
None of the autoscaling solutions presented in the post address the database issue either.
I agree that in general the database is always going to be your bottleneck and autoscaling appservers is generally bullshit.
But let’s assume your DB is fine and you do actually need to scale appservers beyond what your provisioned capacity can allow, you can trivially spin up VMs at AWS or your provider of choice, Wireguard to your “VPC” at your bare-metal provider (or just SSH tunnel the ports) and run your appserver binary/container/etc on it.
Hell in a pinch you can do the above with your dev laptop and serve traffic from it. It will work and your customers won’t even be able to tell.
I disagree, NYC was OK but LA and SF specifically, the average quality of food was pretty bad. There were absolutely good options, plenty of them, but on average, just walking around, seeing something that looks good and checking Google Maps reviews (including reading them), had 0 guarantee of the food being of any quality.
Compared to London and Paris where you're guaranteed a good meal in any random place as long as the google reviews are ~3.5+. And even for the lower, the complaints are usually that the service is slow or rude, not that the food is shit tier quality. I mean 4.9 noted places with ramen that tastes like margarine and uses canned vegetables, or where the meat tastes bad.
Wow, my experience is so different. The two most disappointing places I’ve ever been food wise were London and Paris. Spent 2 months in Paris, nothing stuck out except for one home cooked meal at a friend’s
as Tyler Cowen says, Tokyo has the best food in the world, and the best French food in the world.
> Paris has almost no East Asian food and what I did see when trying to find some was all in porcelain trays like slightly fancier Panda Express.
This is simply not true. Are you sure you were in Paris, France? French people absolutely adore East Asian, and especially Japanese cuisine. There are the meh fast food options, yes, but the majority are definitely sushi places. And you have tons of others like Korean, Korean barbecue, more traditional japanese, etc.
The Japanese restaurants are concentrated around rue St Anne; the Chinese ones around rue Volta. But not exclusively.
As a fun exercise, Japanese is the highest represented origin (country) category in the Michelin guide. After that is Italian with slightly less. (The first category is "modern" which can mean anything).
reply