No. I'd say their algorithm is just as bad as everyone else's since it's designed to promote ragebait (inasmuch as their goal is to increase the odds of "positive engagement" in the form of likes, retweets, and replies, and ragebait happens to maximize that). Engaged audiences stay longer and see more ads. That this results in the utter destruction of online discourse is, of course, not something that they care about. Consequently, I don't really use Twitter for anything other than staying in touch with friends... at least, those I'm not already in contact with through chat rooms of some kind (where we can share whatever we like without some Twat inserting themselves and deciding we didn't really want to share that and oh hey, how about this obnoxious thing Trump did the other day???).
I haven’t touched much modern JS. I have done some TypeScript in the last few years but of course that’s a transpiled thing. I think TS isn’t too bad to write functional stuff though I have no idea if people actually enjoy reading my code.
As an European, I wish you were right, but I’m afraid you aren’t.
The EU would use public funding to build some sort of Google alternative and it would take ages, would be mediocre and most money would go to waste. Instead of incentivising entrepreneurship, which is what they probably should do.
We live very well in the EU. We don’t have to have millions in savings in order to retire. Strong worker protection. Plenty of time off. Low crime rates. Most people fantasise with becoming rich, but as in, “I had a rich aunt that I didn’t even meet in my life and I was the sole heir” or “I won the lottery”, not as in “I grinded for the best 10 years of my life working 100 hours per week before I sold my company” that seems more prevalent in the US. Ordinary people here are super happy if they can buy a small place to live (not a humongous house) even if it takes 25 years to pay it in full, then finish work at 5 and take their kids to the park and have dinner at some restaurant on Saturday.
OTOH: I think the current US administration is the best think that could happen to the EU, a big wake up call. Suddenly there’s money to invest in Defense and that kind of thing.
Also, hopefully LLMs will diminish Google’s importance, and as long as there’s competitive models not from the US (Mistral, DeepSeek) we might be fine. But Google holds all the cards (data). With stuff like the Harvard animosity they might even stop attracting all the foreign talent.
Apple? There’s Samsung for phones at least. Amazon? They’ve become a Temu/Aliexpress. Facebook… huge win if they stopped doing business in Europe. MS? This is the year of Linux in desktop?
The Cloud is one of those things where the EU could build something competitive/alternative just with public funding. All running on Linux, of course.
> The EU would use public funding to build some sort of Google alternative and it would take ages, would be mediocre and most money would go to waste. Instead of incentivising entrepreneurship, which is what they probably should do.
Something like web search is basicallly part of a modern digital infrastructure. We don't want entrepreneurship in water or energy supply, I don't think we should rely on it in web search, because it will inevitably end up chasing profits over everything else.
Markets are good at driving cost down via competition, but once you reach a monopoly steady state there's not much left to optimize. I think a search utility would work just fine. The main barrier to entry is the huge storage and processing resources that are needed to make a good index. Google contains all the information out how to scale like that too.
> I think the current US administration is the best think that could happen to the EU, a big wake up call. Suddenly there’s money to invest in Defense and that kind of thing.
Not to derail the conversation, but IMO the current US administration isn’t a wake up call. It’s a temper tantrum by people who understand that the US isn’t as relatively wealthy to the rest of the world as it was after WW2 but don’t understand why. If some of the thrash accidentally improves the West’s defensive posture or spending that’s good but there is no coherent plan of why things need to be changed.
> Most people fantasise with becoming rich, but as in, “I had a rich aunt that I didn’t even meet in my life and I was the sole heir” or “I won the lottery”, not as in “I grinded for the best 10 years of my life working 100 hours per week before I sold my company” that seems more prevalent in the US.
I promise you that within the US, each of those first two fantasies is more popular than the third one.
>We live very well in the EU. We don’t have to have millions in savings in order to retire. Strong worker protection. Plenty of time off. Low crime rates.
We'll see how that pans out when the baby boomers finish retiring. Europe ate it's children to feed the retirees.
Yes the voices echo all around the globe of what's happening in the US. Surprisingly the only place it isn't heard is in some places of US.
But I still feel like some points raised by the gp might be right. And I was laughing a little thinking that someone critizing the EU already makes you consider them as an american.
Like its just funny.
Also, I feel like every country has problems but countries should honestly first and foremost try to stay away from corruption and the billionaires/rich people's influence in general and try to be impartial. I do think that EU might be good in that but still, I sometimes wonder if this all might be a facade in the sense that EU wants to work and they want to show something for it and so that's why they are fining google only almost 3 billion$. Like maybe my trust in political systems is a little too faded seeing US instititutions erode in days (speaking as non american but I really admired american politics, not anymore)
> And I was laughing a little thinking that someone critizing the EU already makes you consider them as an american.
Look at the comments in this post. The always pro-privacy, anti-ads HN suddenly moaning about this fine. Now that’s super funny and worth of a good laugh. Of course it’s an America vs EU thing, patriotism trumps (no pun intended) all else.
On a more serious note as much as you can be when you realize that discussions aren't happening in good faith and that biases like nations come...
I think that why nationalism/patriotism works is that the state has a monopoly over (legal) violence / laws in general. But the only way that might work is if we believe into them & thus nations have massive wheels of (propaganda?) or whatever it might take to convince the masses to be patriotic.
I feel like everyone all around the world is kinda the same man. We are homo sapiens. Nations shouldn't define us or the way we interact in an ideal world but I feel like a hypocrite when I myself defend my nation sometimes. I generally prefer decentralization to the point that we might take pride in our nations but we don't get influenced by it because the bigger the nation, the larger its influence/propaganda.
I feel like switzerland might be a good example in the sense that I have heard that there are people who don't even remember the (president/prime minister's?) name while working fine. I wonder if the whole world could essentially agree on international laws while being decentralized.
I just feel like that most of us are puppets and very few puppeteers in this world essentially controlling us / manipulating us into doing things that we generally wouldn't do.
A lot of the things you listed are already partially true of the EU. I wouldn't exactly call it fully decentralized, but I doubt many Europeans know by heart the name of the current EU president (I don't even know the proper title of the office) and they would fervently reject the notion of Europe as a single “nation”. Despite, I see more and more people (esp. in threads like this one) describing themselves as “European” rather than their nationalities and crediting EU laws and institutions ahead of their national governments.
I find this trend encouraging and I hope that one day we can see ourselves as humans ahead of any artificial groupings we sort ourselves into.
They also started to support censorship and government approved social media content.
See the recent thread about Nepal banning many apps and the comments are full of people saying that the EU should do the same or require that the content be even more moderated on all the platforms.
It's very sad to see.
I come here for the hacker news but it seems we are being overrun by a new kind of people who love when the EU intervenes to "regulate" the markets and fully believe that the EU is "pro" privacy (TM) and can't wait to impose new regulations all the while it's actually working to undermine encryption for everyone in Europe.
I guess the old saying is true, War is peace, freedom is slavery, ignorance is strength.
Question for the experts: a few years back (even before covid times?) I was tasked with building a news aggregator. Universal Sentence Encoder was new, and we didn’t even have BERT back then. It felt magical (at least as a regular software dev) seeing how the cosine similarity score was heavily correlated with how similar (from a semantic standpoint) two given pieces of text were. That plus some clustering algorithm got the job done.
A few months ago I happened to play with OpenAI’s embeddings model (can’t remember which ones) and I was shocked to see that the cosine similarity of most texts was super close, even if the texts had nothing in common. It’s like the wide 0-1 range that USE (and later BERT) were giving me was compressed to perhaps a 0.2 one. Why is that? Does it mean those embeddings are not great for semantic similarity?
It's likely because the definition of "similar" varies, and it doesn't necessarily mean semantic similarity. Depending on how the embedding model was trained, just texts with a similar format/syntax are indeed "similar" on that axis.
The absolute value of cosine similarity isn't critical (just the order when comparing multiple candidates), but if you finetune an embeddings model for a specific domain, the model will give a wider range of cosine similarity since it can learn which attributes specifically are similar/dissimilar.
That’s not necessarily true. If the embedding model hasn’t been trained on data you care about, then similarity might be dominated by features you don’t care about. Maybe you want documents that reference pizza toppings, but the embedding similarity might actually be dominated by the tone, word complexity, and the use of em dashes as compared to your prompt. That means the relative ordering might not turn out the way you want.
I was reading somewhere that the BERT and USE style were "big-symantic space" designed to 0.0-1.0 so that things unrelated would be close to 0.0, and are classifiers.
But now, like the OpenAI embedding you're talking about the embedding are constrained, trained for retrieval in mind. The pairs are ordered closer, easier to search.
* Limiting the number of QCs (Quick Charges / DC Fast Charge), as this heats up the uncooled Leaf battery, degrading it slightly each time, especially on hotter days
* Keeping the charge between 50-80% when manageable
* Charging up to 100% at least once a month, and letting it 'top off' to rebalance the pack for at least a few hours afterwards
* Not driving like a maniac, despite having more torque in this car than I've ever had in any of my previous cars
This kind of thing (minus the driving like a maniac bit) is what puts me off EVs. I guess it's unavoidable? My experience with laptop and phone batteries (holding much less charge pretty quickly) doesn't help. My phone (iPhone 12) says battery health is at 81% but it doesn't feel like it so I'm not sure I'd trust that Leaf saying it's got 93%.
This is completely avoided and not my experience.
This is only because the leaf is not actively cooled.
Most ev do not suffer from such difficult management of battery and have a computer dedicated to cool / heat and keep battery in healthy temps.
They do degrade over time but very, very slowly. Absolutely not like phones.
Mine has 25k miles and zero degradation yet.
A VW e-Up(and its siblings, Skoda Citigo EV and Seat Mii Electric) all have passively cooled batteries but owners don't report much if any battery degradation even on the first gen models which are over 10 years old now. I can only assume it's the battery chemistry or cell composition compared to the Leaf. Our own is 4 years old and I haven't noticed any range loss compared to when it was new.
Could be due to American climate and Leafs getting sold heavily in America. Jeff lives in a place hotter and sunnier than Málaga. Cars can roast in their huge unshaded parking lots with black tarmac.
Not a battery expert, but I did recently look at using old leaf battery cells to build home battery. Their modules are 2s2p which make them impossible to balance.
Rivian
I think it’s eating the extra buffer but that also means that the worse degradation that happens in first year is « free ».
Then stats are a few percents per 100k miles….
There are articles about stats on long term degradation - it’s a non issue. Buy used if you can in USA - people perceive them as degraded so price is cheap but they don’t degrade fast
Funny, how 2 years old car is really old. For combustion engines, old is like 15+?
EVs will have the same problems as mobile phones. Maybe manufactures want that. But software will define the age of your car and I don’t like it. What if the car requires internet connection and the company dies?
For practical purposes it is a 15 year old design; it had a minor update 8 years ago, but was not brought up to date with the norms of the time (in particular, the passively cooled battery).
This was the very first mass-market electric car, with only minor tweaks. It's not surprising that it's a bit rough around the edges. That's part of why they're cheap second-hand (along with the fact that the 3rd gen, which actually is a proper redesign, is coming out this year).
I don't think that we can use the general "old" here. The statement should have been more explicit and include the design of the battery.
If the model gets even minor updates, manufacturing is happening right now, the car is either good enough or new enough in general, and we cannot use old as negative quality.
Electric motors existed before the combustion engine, and people keep talking about "rapidly evolving area", while the only thing that is rapidly evolving and specific to EVs is the power source. A battery, to be precise. It is all about battery, and nothing else.
The problems with this car are all around the battery pack (though honestly these problems are a little exaggerated). The battery pack is, for practical purposes, a 15 year old design. This car is no longer made. When it was being made, it was the only electric car you could buy with such an old battery pack design.
Buying a 2nd-gen Leaf in 2025 would be a bit silly, unless you were getting a major discount.
I have a 2025 2nd gen on lease because the deals in Colorado were insane and the previous cost of my gas basically pays for the lease + insurance + electricity. That said, unless they give me a killer deal with a major discount at the end of the lease it's going back to the dealer. I love the car and it's been perfect for around town and the region but with 70% of the lease left they want $18000 for buy out. For that price you're well within a nice used EV with active thermal management, modern charge connectors not requiring an expensive adapter, etc...
There's probably going to be a surplus of off-lease Leafs coming up over the next one to three years which might make a great deal for people who can charge at home.
Yeah, I think this is the only way the could sell them, at this point, with the 3rd gen about to come out. Though, where did you even buy it? I vaguely thought they had stopped making them entirely last year in preparation for the third gen (and if I go to Nissan's website it just offers me a third-gen).
ZE1 Leaf(2017) was electronically a big minor update over ZE0(2010). IIRC so much so that ZE1 battery packs almost work on ZE0 body with only minor hack efforts.
So they're basically 15 years old, technologically older than the Model S. Windows 7 was 1 year old when its basic systems shipped.
To add to what others said: back in 2017 40kWh was the standard, now people scoff at 55kWh. Cars also charge faster, which greatly increases their highway driving potential.
> What if the car requires internet connection and the company dies?
That's not a problem specific to EVs and with Chinese combustion car brands coming and going all the time, it's obvious.
Limiting QC events is easy enough. A daily commuter vehicle can easily trickle charge over night, and even the measly 40 kWh old used Leaf can get you 200km a charge - assuming an average commute of 20 km for Germany, that's a whole work week worth of battery life. The only time you as an average person actually "need" QC is for the yearly vacation road trip, but as the author writes, renting an ICE or chungus electric vehicle for that occasion is way more cost effective. Admitted: if you don't have access to trickle charge at either the workplace or your home, the situation looks different.
As for the battery health rating, it's easy enough to measure. Go on the highway, keep it at 80 km/h straight and note how much range you get out of it. In practical commute settings, range will be longer than that anyway due to regenerative braking in all that start-stop-start-stop dance.
This is the thing that ICE people need to internalise.
You THINK you're driving long distances every day and you THINK charging is a massive hassle where you have to drive to a Charging Site and wait for the car to charge for HOURS.
When in reality you plug in at home and have a full battery every morning.
And when the infrastructure is properly built (yay Finland), you can get a week's charge when you're getting groceries as the shop has multiple 100kW chargers along with a fleet of Level 2 (22kW) chargers.
The only times I need to actively think about charging are over 200-250km day trips (my old Ioniq EV has a WLTP of 300km on a warm summer day). And even then the kids an dogs need a bathroom break anyway and I need to walk around a bit to freshen up. 20-30 minutes gets us 100-150km of charge (old car, slow charging) and we're off again.
I had a leaf on demonstration for 24 hours. I was impressed with it. I drove the 50 miles to work and back, and put it on slow charge knowing I had 16 hours until I needed it. No problem. When I came to drive it in the morning it hadn't charged. It was actually a problem with the plug in the garage, it must have gone off a few minutes after I plugged it in. Now what? Wait 16 more hours?
A couple of weeks later I went on holiday. I wouldn't be able to charge where I was stopping, and there were no chargers within 15 miles. I kept my diesel.
> It was actually a problem with the plug in the garage, it must have gone off a few minutes after I plugged it in. Now what? Wait 16 more hours?
So what, that's something you find out once that your electric installation is shoddy enough to most likely be a fire risk (because the charger plug has a thermal protection built in!), fix it and then you won't have that problem again. And as the Leaf should have 150 miles worth of range, you still should have way more than enough range on the battery to do two days worth of commute even if you suddenly find out the trickle charge didn't work.
Renting a car for vacations / road trips is expensive. sure you can but expect to pay $100 per day. Also expect to arrive and be told they can't fill your reservation as they are out of cars.
> Renting a car for vacations / road trips is expensive. sure you can but expect to pay $100 per day.
Indeed... but now, think of the price difference between a small-ish commute EV and a chungus EV or ICE. Easily tens of thousands of dollars, that's a lot of days worth of rent.
> Also expect to arrive and be told they can't fill your reservation as they are out of cars.
That's extremely fucking rare to happen. In the eventuality your reservation can't be filled, you'll usually get upgraded for free. Personally, got upgraded from a small VW Golf class to a VW Phaeton once, plus a day for free. And automatic, no stick shift like the Golf.
Over the life of a car tens of thousands of dollars isn't that hard to build up though if the car cannot do anything. And there is value in having the vehicle that does what you need when the inspiration strikes.
Not having any rental cars is very common in my experience. I'm often renting in smaller cities though (like the rental car place just a couple miles from my house), I've never heard of problems in big tourist destinations or large cities.
Seems expensive but in the grand scheme of a road trip that is only like a 33% increase in price when you factor in a motel room for $100 a night, maybe $100 spend on gas a day filling up twice, $100 on three meals coffee and snacks for one. Give or take of course.
That is probably why the american road trip is a dying animal. Past a days drive you are coming way behind just flying straight there.
I thought about this too - own a small electric vehicle for 90% of the family needs and rent a bigger SUV when the need arises.
Until you do the math and realize that the 3-4 annual trips of multiple days would end up costing thousands of dollars in rental fees per year.
Plus the usual inconveniences around renting.
> Until you do the math and realize that the 3-4 annual trips of multiple days would end up costing thousands of dollars in rental fees per year.
First, 3-4 annual multi-day trips that go for longer than 300 km? If one has that amount of disposable income to afford that, go for whatever the biggest Tesla is and use Superchargers along the route, even drives so long they're a safety issue on its own due to fatigue don't get that much longer due to charging because kids will need to go to the toilet every so often even with an ICE.
As for the rental fees: here in Germany, I just checked - a Mercedes Benz Vito, so up to 8 people (or 6 people plus a ton of luggage), that's 50€ a day here. Crossing four digits takes 20 days of rental, that's a lot of vacation time even by European standards.
For that kind of distance I take a train, Munich-Dortmund doesn't take that long. As for holidays, Croatia by bus or by night train to get around horrid traffic, extortion level road tolls and dumbass border controls.
Those were comporomises Nissan made in 2010s to build an EV that will be under $30k new in 2025. Not all of it are fundamental limitations of Li-ion BEV technology.
Tangential trivia: BYD Dolphin Baseline is 20k new before subsidies in some places.
I don't think about any of this and never have. My 2022 Model Y has 60,000 miles on it and the battery has only lost a negligible amount of health/range since I bought it.
My understanding is that 2 & 3 are problems with the original (20kWh) Leaf model. Which possibly led to a rumour about the 40kWh models, that the OP has.
My Leaf from 2019 has 100% battery health and I always charge it to 100%. I almost never use fast charging though since it is a commute kind of car.
Is that the health reported by the car dashboard, or by LeafSpy? The dash has some wiggle room in its 1-10 display an IIRC it will display 100% until the SoH is below like 90%...
Obviously they are reporting the dashboard number. The first bar drops at 85%. For a 2019, it is probably getting close to dropping its first dashboard bar.
After 25 years in the industry I should know that there is always someone that have read more forums and researched more :)
Either way, the car is 6 years old. At this rate the battery will still be pretty fine when it is 15 years old and ready for retirement.
Never charging it above 80% is not worth the hassle. It is like never using your left hand for the fear of hurting your left hand. You are crippling your car by overthinking.
The performance/battery life of a phone can be seriously affected by the software (each new iOS being built for newer CPUs), whereas this is less of a factor for cars.
If you think this is what puts you off EVs, then you haven't read the horror stories from EV Clinic.
TL;DR is many EVs and hybrids (especially European ones) have tonnes of design faults with e-motors and power-electronics that not only make them ticking time bombs(not in the explosive sense) out of warranty, but also have malicious DRM making third party parts impossible to source, and repairs difficult and eye watering expensive even if theoretically EVs should be more reliable on paper than ICE cars.
Maybe the EU should focus more on EV/auto repairability regulations instead of smartphones and USB-C widgets.
Luckily EV Clinic is working hard on breaking the DRM and reverse engineering parts to make and sell aftermarket ones, but this shouldn't be needed in the first place if the OEMs weren't so bad at design, greedy and hostile to consumers and aftermarket repairs.
Seriously, we need regulations here ASAP. The free market doesn't work here for the consumer when OEMs all do the same anti consumer things.
The DRM thing unfortunately has nothing to do with EVs and everything to do with the computerization of cars. All cars have gotten more and more computerized and harder to repair with more DRM. EVs arguably fundamentally require heavy computerization more than ICE cars, but you're not gonna avoid it in a brand new diesel car either.
The EU did put in regulations, but they did the opposite and essentially mandated DRM. They want to prevent owners from turning off nanny devices or overriding pollution controls.
Hybrids are a different breed, they're very very bad. Don't ever buy one out of warranty and be REALLY sure what the warranty covers.
FYI: Mercedes Benz hybrids have a full(*) battery warranty. What's the asterisk you ask? The warranty only covers repairs up to the car's current value. Which has plummeted over the few years you've owned it and the cost of the battery is absolutely ridiculous. Like "buy 3 used Leafs" ridiculous.
Most hybrid batteries are also blobs of molten plastic and silicone that can't be repaired at all. EV batteries are constructed of semi-standard cells and can be opened and repaired piecemeal by specialist shops. Official repairs usually just swap the whole thing - again for a massive cost.
At a standard dose, paracetamol slightly reduces fever,[13][17][18] though it is inferior to ibuprofen in that respect[19]
[13] Warwick C (November 2008). "Paracetamol and fever management". J R Soc Promot Health. 128 (6): 320–323. doi:10.1177/1466424008092794. PMID 19058473. S2CID 25702228.
[17] Chiumello D, Gotti M, Vergani G (April 2017). "Paracetamol in fever in critically ill patients-an update". J Crit Care. 38: 245–252. doi:10.1016/j.jcrc.2016.10.021. PMID 27992852. S2CID 5815020.
[18] de Martino M, Chiarugi A (December 2015). "Recent Advances in Pediatric Use of Oral Paracetamol in Fever and Pain Management". Pain Ther. 4 (2): 149–68. doi:10.1007/s40122-015-0040-z. PMC 4676765. PMID 26518691.
[19] Pierce CA, Voss B (March 2010). "Efficacy and safety of ibuprofen and acetaminophen in children and adults: a meta-analysis and qualitative review". Ann Pharmacother. 44 (3): 489–506. doi:10.1345/aph.1M332. PMID 20150507. S2CID 44669940.
Errm, No :-) I meant bars as in benchmarks, often rather meaningless, because within the range of statistic noise.
For instance, something having 100.200 points in one config, in another 100.220, with the bars/scales distorted to make that difference seem much larger.
OpenAI recently played a bit too hard with their GPT-5 announcement. Two bars with the same height but wildly different values, things like that. Such a lack of subtlety that their claim it was accidental is actually almost believable.
reply