Together with next-generation ML accelerators in the CPU, the high-performance GPU, and higher-bandwidth unified memory, the Neural Engine makes M4 an outrageously powerful chip for AI.
In case it is not abundantly clear by now: Apple's AI strategy is to put inference (and longer term even learning) on edge devices. This is completely coherent with their privacy-first strategy (which would be at odds with sending data up to the cloud for processing).
Processing data at the edge also makes for the best possible user experience because of the complete independence of network connectivity and hence minimal latency.
If (and that's a big if) they keep their APIs open to run any kind of AI workload on their chips it's a strategy that I personally really really welcome as I don't want the AI future to be centralised in the hands of a few powerful cloud providers.
All: Please don't use HN for ideological battle. There are too many low-quality/predictable comments here. We want curious conversation, not sharp recitals.
I know it's hard when the topic is itself an ideological battle, but that's a good time to review the site guidelines, including this one: "Comments should get more thoughtful and substantive, not less, as a topic gets more divisive."
"and block its websites." So this keeps Israelis from reading Al Jazeera.
Now that's new. Israel started Internet censorship in 2017.[1] Initially it was limited to "terror group websites, online illegal gambling, prostitution services, hard drug sales". At the time, "due to warnings from rights groups that the law poses a slippery slope toward additional censorship, the final version of the legislation dictates that rights groups may appeal the decisions."
Then, in 2021, there was the "Facebook bill", authorizing very broad censorship.[2] That does not seem to have passed. It was first proposed in 2016, almost passed in 2018 [3], tried in 2021, and tried again in 2022. It doesn't seem to have passed.
But something new happened recently. Wikipedia has a note at Censorship in Israel: "This article needs to be updated. The reason given is: New ban issued by the knesset on foreign media channels. Please help update this article to reflect recent events or newly available information. (April 2024)"[4]
The Knesset gave the government the authority to ban foreign media on April 1, 2024.[5]
This isn't just about preventing outside media from reporting from Israel. It keeps Israelis from viewing media the government doesn't like. Haarez has good coverage.[6]
The US White House press secretary issued a weak statement condemning Israel's action, but it was on April 1st and the costumed Easter Bunny overshadowed that statement.[7]
"With these improvements to the CPU and GPU, M4 maintains Apple silicon’s industry-leading performance per watt. M4 can deliver the same performance as M2 using just half the power. And compared with the latest PC chip in a thin and light laptop, M4 can deliver the same performance using just a fourth of the power."
That's an incredible improvement in just a few years. I wonder how much of that is Apple engineering and how much is TSMC improving their 3nm process.
In the spirit of trying to learn something from the community: while I can certainly understand the rationale and goals behind DEI programs (many of which I agree with, others not), I honestly can't understand these "DEI statements" at all. They always seemed very "1984" to me, and almost designed to engender resentment in a way that would ultimately backfire. So perhaps I'm in a like-minded echo chamber, but is there anyone that actually defends these DEI statements with a coherent argument, or can you point me to one online? If so, I'd honestly love to hear it, and I mean this quite genuinely. I did some googling beforehand and found loads of "how to write a good DEI statement" articles, but literally every single one of them just took it at face value that these were a good thing to begin with (or, perhaps in their defense, that "academic jobs require it", so you better learn how to write one in any case).
Stepping back, the high-order bit here is an ML method is beating physically-based methods for accurately predicting the world.
What happens when the best methods for computational fluid dynamics, molecular dynamics, nuclear physics are all uninterpretable ML models? Does this decouple progress from our current understanding of the scientific process - moving to better and better models of the world without human-interpretable theories and mathematical models / explanations? Is that even iteratively sustainable in the way that scientific progress has proven to be?
Apple usually massively exaggerates their tech spec comparison - is it REALLY half the power use of all times (so we'll get double the battery life) or is it half the power use in some scenarios (so we'll get like... 15% more battery life total) ?
It's kind of odd to me (as someone who used to live there at its latest boom time) that nobody talks about Kansas City when it comes to this topic.
From the ~70's until the early 2010's Kansas City's downtown was in a similar "doom loop" of crime, undevelopment, decaying historic buildings, etc... In that city 75% of the metro lives in suburbs, drives in to downtown for work and promptly leaves. Until about 2012 or so. Urban redevelopment kicked in, adding (free!) transit, boosting retail, arts district events, a new stadium, and crucially - *massive office to housing conversion projects*.
There are tons of success stories like the historic Fidelity Tower at 909 Walnut (https://en.wikipedia.org/wiki/909_Walnut), a huge 35-story tower that sat vacant (creepy) for the better part of a decade and is now home to 159 units. Ditto with the Power & Light Building (https://en.wikipedia.org/wiki/Kansas_City_Power_and_Light_Bu...) (36 stories) - largely vacant for the better part of 20 years and now home to nearly 300 units. I could go on, every block has similar projects of 100+ year old buildings of nontrivial sizes that are now super unique apartments. I myself lived in the 30-story Commerce Tower (https://en.wikipedia.org/wiki/Commerce_Tower) for a while and it was incredibly cheap to do so (~$1100/month for 750sqft 1 bed on the 14th floor), I had a 10 minute commute by foot to my office, it was awesome. Even the more squat, broad midsize banking buildings have had major success with residential conversions.
These kinds of conversions have been proven out when there is willpower to do so at the city level - people will move in and prices typically get competitive fast if done at scale. I've lived in SF for 4 years now and I'm convinced its a policy problem not an economic problem.
Take a closer look at how much homework children have, and how much assigned, mandatory reading children have.
I went to a private "college prep" high school, and the amount of assigned, mandatory reading was insane. I spent so much time reading assigned novels that I just didn't have time to read for pleasure. (And why read for pleasure if I already spent 30+ minutes reading a boring / awful novel as part of my homework?)
Once I was out of college, I rediscovered reading.
Traefik is pretty cool, but suffers from the same, terrible problem of Ansible: there is a lot of documentation, and a lot of words written, yet you can never find anything you need.
I have used it since v1 and I routinely get lost in their docs, and get immensely frustrated. I have been using Caddy for smaller projects simply because its documentation is not as terrible (though not great by any stretch)
Technical writers: documentation by example is good only for newbies skimming through. People familiar with your product need a reference and exhaustive lists, not explanation for different fields spread over 10 tutorial pages. Focus on those that use the product day in and day out, not solely on the "onboarding" procedure.
This is my pet peeve and the reason why I hate using Ansible so damn much, and Traefik to a lesser extent.
> In case it is not abundantly clear by now: Apple's AI strategy is to put inference (and longer term even learning) on edge devices. This is completely coherent with their privacy-first strategy (which would be at odds with sending data up to the cloud for processing).
Their primary business goal is to sell hardware. Yes, they’ve diversified into services and being a shopping mall for all, but it is about selling luxury hardware.
The promise of privacy is one way in which they position themselves, but I would not bet the bank on that being true forever.
All: if you're about to comment in this thread, please review https://news.ycombinator.com/newsguidelines.html and make sure your post is in the intended spirit of the site. If it isn't, please edit it until it is; or simply remember that the internet is usually wrong and refrain from posting.
The intended spirit is curious, respectful conversation in which we learn from each other. Yes, that is hard when emotions run strong, but hard != impossible, and it's what the site rules ask: "Comments should get more thoughtful and substantive, not less, as a topic gets more divisive."
Recently I discovered that many coffee shops, maybe half in my sampling of a couple dozen in different cities, are selling cold coffee (brewed hot, then refrigerated) under the name cold brew, and even the ones that actually cold-brew them seem to be under the impression that it needs to be served cold. I was laughed at in one hipster joint for asking for a steamed or warmed cold-brew, and another one initially refused my request to warm it up saying that would make the coffee extremely sour. (It didn’t) Oh, and at least one other, maybe two, said they couldn’t warm cold brew (in view of both a steamer and microwave) or would have to charge extra (while someone’s cheaper latte was being steamed).
Reading the paper, it’s not clear whether their cold brew has lower acidity (higher pH) than the same coffee hot brewed. It does say that the sonic-brew has the same pH as the normal long-steep cold brew. I’m also curious if this cavitation/sonication brewing process is basically agitating the coffee, or doing something different, and how different it is from manually agitating a cold brew compared to letting it sit still for hours.
The bridge is assembled over 2 nights at a motorway exit (so traffic can bypass it by driving off and immediately back on to the road). During night 1 the two end ramps are assembled and attached together to make a short bridge. During night 2 the ramps are driven apart, the central section is built to reach the full length and the entire structure is driven to the final location.
The entire length is 236 meters long providing a working length of 100 meters underneath. The assembled bridge can flex slightly at the joins between sections, and has a turning radius of 2 kilometers.
Private equity is akin to toxic poison for everyone who isn't one of the (very few) general partners who bask in obscene amounts of scrooge-mcduck money. Think of making a cool $30 million+ a quarter, per partner -- thanks real estate (and apparently, emergency medical services) "investments"! This is a direct funnel to syphon money from the workers to the ruling class. That's even more cash than the top brass at FAAGs make.
How can this phenomenon be reigned in and sorted in the medium to long-term?
It's a loss for everyone (99.9999%) except the miniscule quantity of winners (0.0001%, yes, that's 1 in 1,000,000 - there aren't that many large-scale P/E firms). Housing owned by P/E couldn't be more soul sucking. Scratch that, this applies to every case I'm aware of where P/E firms own a thing.
Anecdote: Last month my dear old dog was suddenly very ill and I spent $12k on emergency surgery to save his sweet life. No regrets, I'm grateful he was fixable. But in the waiting room I did the conservative math for what the SAGE Emergency Veterinary Hospital in Redwood City, CA makes in a year (before the comparatively small overhead, tax-deductible expenses) and they are pulling in at least $30 million in this single location. That's.. a lot, for many 20 total staff. Also they charged me $500 for each follow-up visit, which is way more than my real life surgeon charged me earlier this year! (Even including insurance, he didn't charge much for the follow-ups to make sure I survived.) Damn.
For the dog lovers out there: Doggo is alive and well, can maybe last another year or two we hope.
> Conventional wisdom is that only certain office buildings can be converted to housing.
If you can buy the building cheap enough, conventional wisdom can be thrown out the window.
I once toured a building that had been converted from an old warehouse to residential. Huge floorplate. They had built the condo units around the edges, created a hallway, and then the inside was divided up into "storage" spaces. Each condo owned the space directly across the hallway. They were very large, and people had transformed them into offices, arcades, workshops, playrooms, theaters, etc. You could do just about anything you wanted with the space, and because it wasn't "living space" you didn't have to worry so much about noise and the property taxes were lower than they otherwise would have been.
You can't sell it for a price that includes that space as "living space" either. Which goes back to the point - if you can buy the building cheap enough, you can make anything work.
So next the attackers playing the long game will just set out to develop the next great everybody-uses-it open-source library, so they control it from inception?
Great that we'll finally get state-sponsored open-source development :D
If you're a scientist who works in protein folding (or one of those other areas) and strongly believe that science's goal is to produce falsifiable hypotheses, these new approaches will be extremely depressing, especially if you aren't proficient enough with ML to reproduce this work in your own hands.
If you're a scientist who accepts that probabilist models beat interpretable ones (articulated well here: https://norvig.com/chomsky.html), then you'll be quite happy because this is yet another validation of the value of statistical approaches in moving our ability to predict the universe forward.
If you're the sort of person who believes that human brains are capable of understanding the "why" of how things work in all its true detail, you'll find this an interesting challenge- can we actually interpret these models, or are human brains too feeble to understand complex systems without sophisticated models?
If you're the sort of person who likes simple models with as few parameters as possible, you're probably excited because developing more comprehensible or interpretable models that have equivalent predictive ability is a very attractive research subject.
(FWIW, I'm in the camp of "we should simultaneously seek simpler, more interpretable models, while also seeking to improve native human intelligence using computational augmentation")
Over the weekend a coworker was helping someone in our lab prepare data for an important conference talk.
We accidentally ran up a rather large bill because while the EFS storage pricing was simple enough, the usage pricing bit us.
It seems like AWS' entire business model is making the pricing so confusing that you don't know what it will cost until after you've used it. It feels weirdly similar to the US healthcare/insurance situation.
More competition in this space can only be a good thing.
Just a point about Al Jazeera that is worth mentioning. There are two Al Jazeeras. One which is presented to Western audiences and one which is presented to Middle-Eastern audiences. The media and articles are politically aligned in each region. Don't assume what you read in the US/UK/wherever is the same as over there.
In the case of Israel, the Middle-Eastern unit were literally showing videos demanding further uprising against Israel across the region directly from Hamas. Also the entire point of Al Jazeera was for Qatar to provide political influence through media, not as an unbiased news agency.
It's even banned in some Arab countries for being a security risk.
Why would you allow that to continue in your country?
> The US White House press secretary issued a weak statement condemning Israel's action, but it was on April 1st and the costumed Easter Bunny overshadowed that statement.[7]
It really is amazing. We were thiiiiiiis close to having a generation that wasn't addicted to nicotine for the first time in hundreds of years (like, since trade between North America and Europe really became a thing) and these guys pulled this thumb drive that tastes like cotton candy out of nowhere and every young person is just jonesing.
And the regulatory/public opinion environment isn't what it was in the 80s and 90s when we decided to take on cigarettes. Imagine a company being told today that they now have to stop advertising on billboards, can't advertise to kids, and have to pay a fifth of a trillion dollars over 25 years for the harms they caused to Medicaid. Can't, can you?
I'm a maintainer (one of many) of an open source project, and this topic has been on my mind a lot lately as I review PRs.
I am more suspicious of PRs from new contributors by default now. Of course I keep these suspicions to myself, but besides simply reviewing code for all the regular things, I now ask myself "what sort of sneaky thing could they be doing that appears benign on the surface?"
Design doc culture at google is why I left the company.
I wrote a very high level doc shortly after joining the company outlining a fairly trivial task (one that has been performed many times across other product areas).
A colleague virtually took me aside and told me "this is not how we do things here". Which was kind of a wake up call as he told me to make sure to 'evaluate many more ways to get this done' despite the way I outlined literally being a slight variation on the recommended way.
When questioned about why this would be needed he said "it shows broad consideration".
"Fake work" is very much a thing at google, I wish I joined a different team.
IME Apple has always been the most honest when it makes performance claims. LIke when they said an Macbook Air would last 10+ hours and third-party reviewers would get 8-9+ hours. All the while, Dell or HP would claim 19 hours and you'd be lucky to get 2 eg [1].
As for CPU power use, of course that doesn't translate into doubling battery life because there are other components. And yes, it seems the OLED display uses more power so, all in all, battery life seems to be about the same.
I'm interested to see an M3 vs M4 performance comparison in the real world. IIRC the M3 was a questionable upgrade. Some things were better but some weren't.
Overall the M-series SoCs have been an excellent product however.
This article is a bit difficult to read, as it seems to be written with a heavy dose of sarcasm/irony.
I genuinely can't tell what the author is arguing for, as it's extremely difficult to tell if he's quoting things because he agrees or disagrees with them.
My biggest question is: is the author arguing that there should be spaced bollards along literally every sidewalk in the country/world, and around all edges of every parking lot?
If so, it's an interesting idea, but I also can't help but think that would not just be expensive, but also possibly extremely ugly.
I'm curious if there are estimates of both installation cost as well as lives saved and other damage to buildings avoided.
Thank you for saying this. When cold brew first came out, it was promoted as a brewing process that resulted in smoother (I'm guessing lower acidity) tasting coffee. Heating it up seemed natural, and its use in iced coffee seemed simply opportunistic. (In my experience at least).
Then it quickly caught on as a novelty, with nitro et al, and when I tell people I drink cold brew warmed I get looks of confusion or turned up noses.
Article doesn't mention one of the more interesting (to me) aspects which was how feedback was avoided. The solution is elegant: each vocal microphone is doubled, meaning there are two at each position. The phase is inverted on one of them, the singer sings into only one, and both are sent to the speakers via their channel's amp.
The effect of that setup is that only the difference between the two microphones is amplified; common signal in both (i.e. the sound coming out of the speakers) is nulled out, but the difference signal (the voice) makes it through. It apparently wasn't quite perfect but was absolutely a lot better than wailing feedback.
The thing that made it sound so good was that any given speaker only reproduces a single source, but the article touches on that. The mic arrangement I described is simply what makes it possible.
Why are we running these high end CPUs on tablets without the ability to run pro apps like Xcode?
Until I can run Xcode on an iPad (not Swift Playgrounds), it's a pass for me. Hear me out: I don't want to bring both an iPad and Macbook on trips, but I need Xcode. Because of this, I have to pick the Macbook every time. I want an iPad, but the iPad doesn't want me.
In case it is not abundantly clear by now: Apple's AI strategy is to put inference (and longer term even learning) on edge devices. This is completely coherent with their privacy-first strategy (which would be at odds with sending data up to the cloud for processing).
Processing data at the edge also makes for the best possible user experience because of the complete independence of network connectivity and hence minimal latency.
If (and that's a big if) they keep their APIs open to run any kind of AI workload on their chips it's a strategy that I personally really really welcome as I don't want the AI future to be centralised in the hands of a few powerful cloud providers.