Google censors the world, together with Microsoft.
Well - it is time that the rest of the world censors these two corporation. I don't want them to restrict information.
People will find workarounds by the way. This is now a Streisand effect - as people see that Google and Microsoft try to hide information from them, they will now look at this much more closely than before, with more attention.
(Having said that, my bypass strategy is to not use Windows 11 altogether. I don't depend on it, having used Linux since 21 years now, but my machine to the left is actually using Win10, for various reasons, such as that I can fix problems of elderly relatives still using Windows. But I won't use Win11 ever with its recall-spy software. I also don't care that it can be disabled - any corporation that tries to sniff-invade on me, is evil and must be banned.)
Edit: Ok so the video was restored. That was good, but still, we need an alternative here. Google holds WAY too much power via youtube.
> From that point on, Steve would go on to spend lavishly on things that improved the experience, and he would reject—often brutally—any idea that diluted or harmed the experience. ...I’ll go out on a limb and say that uninvited advertising is not normally equated with a better customer experience.
YES!!! SOO much of the Apple user experience has degraded due to this. I can't listen to my own music that I bought on the Music app, without being interrupted asking if I want Apple Music. I open up the Books app to read Winnie the Pooh to my son, and the opening screen has loads of random trashy romances to try to sell me. I go to comfort read Ender's Game, which I did buy though the store a decade ago, and it helpfully "groups" it with the other four (!?) books in that series which I haven't bought, as if to say, "Don't you want to buy these too?" NO! If I want to buy them, I know where to find them!
It is SUCH an unpleasant experience. EVERY time I open the App Store to update some apps, I'm angry that I have to wander past advertising assaults to do it. EVERY time I open the music app to play an old favorite, I'm angry that I have to go past the advertising assault. EVERY time I open up the book app, I'm angry that I have to go past the advertising assault.
I very much doubt the execs understand how much they're damaging the brand for that little bit of extra revenue. The see the extra revenue, but they don't see the lost brand, or the people that switch away. Is it really worth it?
ETA: I don't think it's an exaggeration to say:
Modern iPhones don't come with a music player. They come with a music store, that you happen to be able to put your own music into. But it's not structured to help you play your music, it's structured to sell you what they want to sell you.
Modern iPhones don't come with an e-book app. They come with a book store that you happen to be able to upload some of your own books into. But it's not structured to help you organize and read your books -- even the ones you've bought; it's structured to sell you more books.
Google must have some awful PMs and designers. The worst UX decision I have seen recently is AI auto-dubbing all youtube videos by default with no way to disable this behavior globally. How could you miss that people can be fluent in multiple languages and if I click on a video in a foreign language, I most likely want the original soundtrack? Clearly, the intention was to boost some metric “X users are using this feature” with no regard for the actual user.
As a parent who gave my oldest child a (very used) smartphone just before she turned 14, I would be in favor of making smartphones illegal under age 15 (or some other number, higher or lower I don't care). I'm pretty sure they're worse than cigarettes for the future of humanity.
Many of the comments here are judging PyTorch in hindsight, which is unfair.
When Soumith Chintala co-launched PyTorch, and for many years after, the alternatives for fast, interactive, convenient development were much worse. There was no Jax.
Every single AI researcher I know, including me, who tried PyTorch back then immediately wanted to switch to it, because it was so much better. Andrej Karpathy described what PyTorch felt like back then when he tweeted, on May 2017, "I've been using PyTorch a few months now and I've never felt better. I have more energy. My skin is clearer. My eyesight has improved."[a]
THANK YOU SOUMITH for your hard work over all these years! Your hard work has made a difference for a huge number of people, including many of us here on HN.
We wish you success in your future endeavors, whatever they turn out to be!
Please ignore all the petty criticism.
---
[a] https://x.com/karpathy/status/868178954032513024
I don't see how the yearly tech support I do with my parents at Christmas will not one day converge to an outright ban of the internet. I am now demoing the level of sofistication of AI powered scams, telling them that it is now entirely possible they will get a VIDEO CALL from me that's not actually me asking for God knows what in a very convincing way using my face and voice. I am scared and this close to setting up a secret passphrase in case they need to tell me appart from a clone.
Alongside a password manager and keeping things up to date, using an ad blocker is truly a foundational security practice these days. The big advertising players simply have all of the wrong incentives to control this problem. They could massively reduce the volume of scams advertised on their networks, but it’d be worse for them on two fronts: they’d have to pay for more moderation, and they’d lose billions in revenue in the process. Shoulder surfing while a non-savvy user browses Facebook or YouTube without an ad blocker and engages with obviously fraudulent ads is painful.
There is still the Old Farmer’s Almanac https://www.almanac.com/old-farmers-almanac-233-years-and-st...
As a Chinese user, I can say that many people use Kimi, even though I personally don’t use it much. China’s open-source strategy has many significant effects—not only because it aligns with the spirit of open source. For domestic Chinese companies, it also prevents startups from making reckless investments to develop mediocre models. Instead, everyone is pushed to start from a relatively high baseline. Of course, many small companies in the U.S., Japan, and Europe are also building on Qwen. Kimi is similar: before DeepSeek and others emerged, their model quality was pretty bad. Once the open-source strategy was set, these companies had no choice but to adjust their product lines and development approaches to improve their models.
Moreover, the ultimate competition between models will eventually become a competition over energy. China’s open-source models have major advantages in energy consumption, and China itself has a huge advantage in energy resources. They may not necessarily outperform the U.S., but they probably won’t fall too far behind either.
On the plus side, Troy can save a lot of DB space now. Instead of storing which emails have been compromised at this point he can replace that with just
def email_compromised(email):
return TrueThere have been enough data breaches at this point that I'm sure all my info has been exposed multiple times (addresses, SSN, telephone number, email, etc). My email is in over a dozen breaches listed on the been pwned site. I've gotten legal letters about breaches from colleges I applied to, job boards I used, and other places that definitely have a good amount of my past personal information. And that's not even counting the "legal" big data /analytics collected from past social media, Internet browsing, and whatever else.
I now use strong passwords stored in bitwarden to try to at least keep on top of that one piece. I'm sure there are unfortunately random old accounts on services I don't use anymore with compromised passwords out there.
Not really sure what if anything can be done at this point. I wish my info wasn't out there but it is.
FBI wants to remind everyone that only US mega-cap companies can scrape the entire internet, not share the data, use it to train AI models and then charge people to use chatbots that use this 'laundered' data. Anybody else attempting to do this is a criminal in their eyes and must be punished.
Related:
"IMPOSING SANCTIONS ON THE INTERNATIONAL CRIMINAL COURT" (white house, feb 2025) https://www.whitehouse.gov/presidential-actions/2025/02/impo...
Microsoft admits in French court it can't keep EU data safe from US authorities (jul 2025) https://news.ycombinator.com/item?id=45822902
As someone who has been the target of an FBI investigation for what was effectively criminal copyright infringement (later arrested and did time in prison), my only takeaway is that this, if anything, should just be a civil suit just like so many other similar cases of copyright issues.
In my personal experience, the priorities of the FBI are typically highly politically motivated. The exceptions are if you’re doing something seriously icky, or doing fraud that deceives people.
For those interested in what’s reported and what actually happens, I’ve made some comments on my case and my experience here: https://prison.josh.mn
> We can not allow the FBI to work for Evil here
Historically speaking I can't see this as even being in the top 100 evil things the FBI has done.
They pardoned the Silk Road drug lord to go after a copyright infringement-lord instead? It's not even in their effective jurisdiction, if this indeed is a Russian national. Don't they have more important Russian crimes to investigate?
I read there was a US government investigation tracking Ukranian children abducted by Russian forces, but supposedly there weren't enough resources [0] to sustain that.
[0] https://www.npr.org/2025/03/19/nx-s1-5333328/trump-admin-cut...
I pay subscriptions to some of these sites and still use archive.is on them because it is a more pleasant reading experience. No auth failures, no annoying popover windows begging me to subscribe to their dumb newsletter. Just the internet equivalent of a static piece of newsprint.
We need to preserve data. The FBI is trying to kill data.
We can not allow the FBI to work for Evil here. I actually think there should be a human right to data. With that I mean, primarily, knowledge, not to data about a single human being as such (e. g. "doxxing" or any such crap - I mean knowledge).
Knowledge itself should become a human right. I understand that the current law is very favourable to mega-corporations milking mankind dry, but the law should also be changed. (I am not anti-business per se, mind you - I just think the law should not become a tool to contain human rights, including access to knowledge and information at all times.)
Wikipedia is somewhat ok, but it also misses a TON of stuff, and unfortunately it only has one primary view, whereas many things need some explanation before one can understand it. When I read up on a (to me) new topic, I try to focus on simple things and master these first. Some wikipedia articles are so complicated that even after staring at them for several minutes, and reading it, I still haven't the slightest clue what this is about. This is also a problem of wikipedia - as so many different people write things, it is sometimes super-hard to understand what wikipedia is trying to convey here.
Ignoring the politics, we have to say that China has done the world collectively as a whole a major service in strategically developing and mass producing super cheap solar panels.
I'm no fan of Cloudflare, but they're completely in the right on that. Infrastructure for blocking websites simply shouldn't exist.
Because if it's allowed to exist, it ends up subsumed by political and corporate interests, and becomes a tool of overreach and abuse. We've seen that happen over and over again.
If US Trade Office can be leveraged to destroy internet censorship efforts in other countries, then so be it.
Doesn't matter. We must keep building more and more technology no matter the cost. Have an idea for a business? Build it. Does your business make the lives of people worse? Doesn't matter, keep pushing. Could some new technology ruin the lives and relationships that people have? Doesn't matter, just build it. We always need more, need to do more. Every experiment is valid, every impulse must be followed. More complexity, more control, more distraction, more outrage, more engagement. Just keep building forever no matter the cost.
No, the people making TUIs in Rust are making TUIs because they love TUIs, and because Ratatui is pretty delightful. The state of GUI frameworks in Rust is irrelevant for this purpose, because even if there existed your ideal of Qt in Rust (putting aside the debatable notion that Qt is some sort of pinnacle of design), the people making TUIs wouldn't care, because that's not what they want to make.
Maybe "open a communication channel" is what Mozilla should have done BEFORE they turned on this thing.
From the article: "It has been working now without our acceptance, without controls, without communications".
This person has been doing volunteer work for a long time, attempting to create a helpful environment. Then suddenly, from above a machine is turned on that shits all over that effort. Makes one feel unwelcome, and unseen...
As someone from the Netherlands I read absolutely 0 condescending or patronising behaviour in this.
It just looks like someone trying to get in contact and help out. I could've written this myself, genuinely trying to help.
This revolution also has a dark side:
People are paying off these devices and then once they have paid them off, they break and people in these areas don’t have the skills or resources to fix them.
This has led to over 250 million of the units lying around broken in peoples homes, leading to solar being one of the fastest growing e-waste streams in the world.
It’s hardly solar punk to sell people cheap crap at a 10x mark up that pretty much immediately breaks once the warranty period is over.
More details for the interested here: https://solar-aid.org/wp-content/uploads/2024/10/State-of-Re...
> Hi Marsf,
> I'm sorry for how you and the Japanese community feel about the MT workflow that we just recently introduced. Would you be interested to hop on a call with us to talk about this further? We want to make sure we trully understand what you're struggling with.
> My timezone is UTC+7, so it should be easier for us to set up time. Let me know how that sound!
What a horribly condescending and patronising response. If I'd received this it would just further vindicate my decision to quit.
I think a fundamental lack of understanding/humility is the core of this conflict along with Mozilla's long and storied history of creating controversies/problems out of thin air.
The Mozilla leadership seems to have a unfortunate tendency to emulate the behaviors of the tech companies that their core Firefox project is often seen as an alternative too.
Firefox is a good browser but is prevented from capitalizing on the skepticism the consumers feel toward the tech sector by Mozilla using the exact same language and dark UI pattern to promote things like pocket that the user-base never asked for, and jump on to the lets enforce the use of AI everywhere that's driving discontent within the proprietary ecosystems, and this is yet another example of this class of behavior from the Mozilla leadership.
"narco-trafficking boats"
There's no public evidence of that though. No trial. It's the same as if we sent the navy to board those boats, put a gun to people's heads and execute them in cold blood.
I'm sorry for how you feel about us kicking you in the balls.
Would you like to hop on a quick call to chat about this further?
Just a quick lil call.
Quick lil ol' callerino.
Hoppity hip hop.
Some people do these type of contribution or charity work not just to do some good but also to feel some autonomy and mastery in a world were much of the regular top down driven drudgery work does not provide much of that feeling. These people are canaries in the coal mine. I expect more people feel a loss of purpose and rise of anxiety and depression in the world.
I'm not really able to understand the finer details but I think I picked up enough to get the broad strokes.
Really though, all I needed to see was the phrase "jump on a quick call" to form an irrationally strong opinion. That phrase instantly warms my entire body with rage.
10+ years in Japan. The message here is much deeper from my perspective. “Let’s jump on the call” is not the solution. The guy was stripped off of his face. I love Japan for being human. Small business bar or restaurant with 3 tables. Not everything should be streamlined for a quick call solution… the process was pushed on his head. Google nemawashi decision making process
This is rich
> We want to make sure we trully understand what you're struggling with.
The post literally starts with a list of grievances. Maybe ask the AI for an executive summary and the key points.
Many reasons, in no particular order:
- It is extremely slow and resource intensive. Opening any link/page takes at least 3 seconds on my fastest computer, but the content is mostly text with images.
- It cannot be used without JS (it used to be at least readable, now only the issue description loads). I want the bug tracker to be readable from Dillo itself. There are several CLI options but those are no better than just storing the issues as files and using my editor.
- It forces users to create an account to interact and it doesn't interoperate with other forges. It is a walled garden owned by a for-profit corporation.
- You need an Internet connection to use it and a good one. Loading the main page of the dillo repo requires 3 MiB of traffic (compressed) This is more than twice the size of a release of Dillo (we use a floppy disk as limit). Loading our index of all opened issues downloads 7.6 KiB (compressed).
- Replying by email mangles the content (there is no Markdown?).
- I cannot add (built-in) dependencies across issues.
I'll probably write some post with more details when we finally consider the migration complete.
It's so insane that they let things go this far. It could have been immediately obvious to those involved that cell phones in class would have immensely negative effects. I mean they talk about a lunch room "quiet enough to hear a pin drop"??
I think I learned half my basic social skills from lunch rooms in school. That time period is probably more important than any of the classes themselves.
As I heard someone say, happiness is your reality minus your expectations.
Smart people see more variables that could be changed, more components that could be modified, and are less likely to accept things as they are. This creates a false sense of ease by which reality could be modified, and thus higher expectations for the world around them.
I suspect this misplaces happiness and contentment, but the two are also very strongly correlated for many people.
The grid is HUGELY expensive, an absolutely massive cost for our electricity. And it would still be expensive in a well-regulated environment where you can quickly and easily get permission to build, without, say, voter ballot propositions illegally blocking a transmission line for years [1]. Here in the US we have a very very poorly regulated environment for adding to our grid, it moves slower than molasses and there are so many parties that have unilateral veto points. The advent of a new transmission route in the US these days is pretty much a miracle event.
Now imagine a world where there's tons of bribes to government officials all along the way to get a grid going (in the US you just need to bribe landowners and hold-outs). Or there's bribes to get a permit for the large centralized electriticy generator. And you have to deal with importing a whole new skill set and trades, on top of importing all the materials, fuel, etc.
Decentralized solar plus batteries is already cheaper than electricity + transmission for me at my home in the US. The only thing stopping me is the permitting hassle or the contractor hassle.
Out in greenfield, solar plus storage is so revolutionary. This is bigger than going straight to mobile phones instead of landlines.
Africa is going to get so much power, and it's all going to be clean, renewable energy. Thanks to all the entrepreneurs and engineers over the past decades that have continuously and steadily improved this technology, it's one of the bright lights of humanity these days.
[1] https://www.utilitydive.com/news/maine-jury-clears-avangrids...
This title is inaccurate. What they are disallowing are users using ChatGPT to offer legal and medical advice to other people. First parties can still use ChatGPT for medical and legal advice for themselves.
Remember: If OpenAI/Google does it for $$$, it's not illegal. If idealists do it for public access, full force of the law.
Information wants to be free. Oblige it. Fools with temporary power trying to extract from the work of others will be a blip in the history books if we make them.
Maintainer here.
We are currently in the process of moving Dillo away from GitHub:
- New website (nginx): https://dillo-browser.org/
- Repositories (C, cgit): https://git.dillo-browser.org/
- Bug tracker (C, buggy): https://bug.dillo-browser.org/
They should survive HN hug.
The CI runs on git hooks and outputs the logs to the web (private for now).
All services are very simple and work without JS, so Dillo can be developed fully within Dillo itself.
During this testing period I will continue to sync the GitHub git repository, but in the future I will probably mark it as archived.
See also:
- https://fosstodon.org/@dillo/114927456382947046
- https://fosstodon.org/@dillo/115307022432139097
Might be useful to ask a different question: What makes people happy?
It's things like relationships, satisfying work, accomplishment. (and many, many more)
Then the real question emerges: How many of those happiness 'sources' are made better by intelligence? What percentage?
Relationships? Seems like no. Work? Also seems like no, lots of work doesn't make use of a high IQ that people enjoy nonetheless. Accomplishment? Strikes me as most likely of the three, but it's also very relative.
And another thought,
Asking why smart people aren't happier is a bit like asking why people who can jump high aren't more empathetic. There's no direct link between the two, you have to dip out to the material conditions. Like: someone who can jump high is fitter > fitter people are healthier > healthier people have more mental time to be empathetic with > people who can jump high are more empathetic. For intelligence, we say smart people are happier. Same thing, happiness is not directly correlated. Instead: Smart people are better able to create the outcomes they want > They select outcomes that make them happy > Their environment makes them happy > Smart people are happier. (These are illustrations of the idea, not actual logical chains or claims.)
I feel like I've been reading this exact same article for the last 15 years.. I find it very difficulty to parse what is real and what is vaporware in the medical breakthroughs community.
I work in rail safety. Two major non-Chinese train companies attempted to merge a few years ago, explicitly to build a company that could compete with China's national company, and provide safer alternatives to state-sponsored cyberhacking of Western rail.
It fell down to an anti-monopoly decision by a single person in the EU ministry, who killed the proposal. Several attempts were made to streamline the merger, but she wouldn't budge.
As a result, CRCC continues to win contracts abroad, largely (it is believed) by undercutting competition. IP theft is known to be one objective of their at-loss or low-profit contracts (I've been involved in fighting that, specifically).
It's hardly a stretch to imagine that having control of the rail in countries that might oppose you militarily is strategically huge.
This article is about busways, but the parallels are obvious.
Many (most?) older retail businesses still use TUIs. They're reliable, consistent, and orders of magnitude faster than GUI systems.
When I worked ar Sherwin Williams, I got good enough with the TUI that customers could rattle off their orders while I punch it into the computer in real time.
It's absolutely crazy that a well designed TUI is so much faster. It turns out that if you never change the UI and every menu item always has the same hotkey, navigating the software becomes muscle memory and your speed is only limited by how fast you can physically push the buttons.
The program had many menu options added and removed over the decades, but the crucial part is that the hotkeys and menu indexes never, ever changed. Once you learn that you can pop into a quick order menu with this specific sequence of five keys, you just automatically open the right menu the moment a customer walks up. No thought, just pure reflex.
UX absolutely peaked with TUIs several decades ago. No graphical interface I've ever seen comes even close to the raw utility and speed of these finely tuned TUIs. There is a very, very good reason that the oldest and wealthiest retail businesses still use this ancient software. It works, and it's staggeringly effective, and any conceivable replacement will only be worse. There simply is no effective way to improve it.
Edit: I will say that these systems take time and effort to learn. You have to commit these UI paths to memory, which isn't too hard, but in order to be maximally effective, you also have to memorize a lot of product metadata. But the key is that it really doesn't take longer than your ordinary training period to become minimally effective. After that, you just pick up the muscle memory as you go. It's pretty analogous to learning touch typing without trying. Your hands just learn where the keys are and after enough time your brain translates words into keystrokes without active thought.
It's a beautiful way to design maximally effective software. We've really lost something very important with the shift to GUI and the shunning of text mode.
It's not as black-and-white as "Brenda good, AI bad". It's much more nuanced than this.
When it comes to (traditional) coding, for the most part, when I program a function to do X, every single time I run that function from now until the heat death of the sun, it will always produce Y. Forever! When it does, we understand why, and when it doesn't, we also can understand why it didn't!
When I use AI to perform X, every single time I run that AI from now until the heat death of the sun it will maybe produce Y. Forever! When it does, we don't understand why, and when it doesn't, we also don't understand why!
We know that Brenda might screw up sometimes but she doesn't run at the speed of light, isn't able to produce a thousand lines of Excel Macro in 3 seconds, doesn't hallucinate (well, let's hope she doesn't), can follow instructions etc. If she does make a mistake, we can find it, fix it, ask her what happened etc. before the damage is too great.
In short: when AI does anything at all, we only have, at best, a rough approximation of why it did it. With Brenda, it only takes a couple of questions to figure it out!
Before anyone says I'm against AI, I love it and am neck-deep in it all day when programming (not vibe-coding!) so I have a full understanding of what I'm getting myself into but I also know its limitations!
No contradiction here:
When we say “machine”, we mean deterministic algorithms and predictable mechanisms.
Generative AI is neither of those things (in theory it is deterministic but not for any practical applications).
If we order by predictability:
Quick Sort > Brenda > Gen AI
The AVHerald is usually the best source for these things, rather than MSM: https://avherald.com/h?article=52f5748f&opt=0
> Ground observers reported the aircraft had been delayed for about two hours for work on the left hand engine (engine #1), the engine #1 separated during the takeoff run, the center engine emitted streaks of flames, the aircraft impacted a UPS warehouse and ploughed through other facilities before coming to rest in a large plume of fire and smoke.
To the authors of the site, please know that your current "Cookiebot by Usercentrics" is old and pretty much illegal. You shouldn't need to click 5 times to "Reject all" if accepting all is one click. Newer versions have a "Deny" button.
I'll give you one reason, among many, it wasn't a failure. It made it illegal to deny people health insurance coverage based on pre existing conditions. That was a big step forward in a broken system to restore some humanity to the system.
Meanwhile in pretty much all other nations you go online to the free website, see your employer contributions already filled in and acknowledge they are correct for the year, add any extra income, check boxes for relevant deductions and you’re done.
> he called for universal health care but once he was elected he started saying it was "too radical"
ACA was the most radical package that could have passed, and it still cost Democrats the Congress.
This line of argument reminds of the folks who complained about Sinema and Manchin. You know what we’d have with a few more Sinemas and Manchins in the party right now? A majority.
The bill that passes is better than the ideal that doesn’t.
Money well spent
https://www.yahoo.com/news/opinion-makers-turbotax-gave-trum...
You guys have it all wrong. There was only one candidate for the dem party, Here's the list:
1) Cuomo. Sexpest who has been accused by many women of some pretty shitty stuff. Also a member of a multi-generational dynasty, which is not good.
2) Mayor Adams. Federally indicted by the Feds. They have a 99% conviction rate. Not because they're corrupt, but because they only go after people who have dome some really egregious, illegal shit.
3) Mamdani. Millennial candidate. No dirt. Other that some stupid stuff he said while he was young, his policies are relatively common sense and middle of the road, and are aimed at leveling the playing field.
Gee, who should I choose? [[said all of NYC today]]
I hope this win signals (to both parties) that voters are receptive and will get engaged when a clear message is presented about cost of living and quality of life issues. Some of which are taken for granted in most other western countries.
I’m no political wonk, and I’m curious what others with more insight might say about his ability to fund and implement his polices.
I’m reminided of Obama and his hopeful message but he was mostly stymied on policy goals. Specifically Obamacare as an example ended up being watered down
If you had told me an article ostensibly about a file format would have me teary-eyed by the end I wouldn't have believed you. This is beautiful, thank you!
Hilariously, this is the second Sam that wants to collect everyones iris's for nefarious purposes
Feels weird to say but I have found using Yandex of all places an excellent search engine for content that get taken down by DMCA requests.
Eg if you want to watch a movie that's not on Netflix using a web stream the search results are far better.
Feels like Google circa 2005.
Anna's archive has already fulfilled G's needs (training Gemini) so now it's time to pretend it never existed ;)
I dislike those black and white takes a lot. It's absolutely true that most startups that just run an EC2 instance will save a lot of cash going to Hetzner, Linode, Digital Ocean or whatever. I do host at Hetzner myself and so do a lot of my clients.
That being said, the cloud does have a lot of advantages:
- You're getting a lot of services readily available. Need offsite backups? A few clicks. Managed database? A few clicks. Multiple AZs? Available in seconds.
- You're not paying up-front costs (vs. investing hundreds of dollars for buying server hardware) and everything is available right now [0]
- Peak-heavy loads can be a lot cheaper. Mostly irrelevant for you average compute load, but things are quite different if you need to train an LLM
- Many services are already certified according to all kinds of standards, which can be very useful depending on your customers
Also, engineering time and time in general can be expensive. If you are a solo entrepreneur or a slow growth company, you have a lot of engineering time for basically free. But in a quick growth or prototyping phase, not to speak of venture funding, things can be quite different. Buying engineering time for >150€/hour can quickly offset a lot of saving [1].
Does this apply to most companies? No. Obviously not. But the cloud is not too expensive - you're paying for stuff you don't need. That's an entirely different kind of error.
[0] Compared to the rack hosting setup described in the post. Hetzner, Linode, etc. do provide multiple AZs with dedicated servers.
[1] Just to be fair, debugging cloud errors can be time consuming, too, and experienced AWS engineers will not be cheaper. But an RDS instance with solid backups-equivalent will usually not amortize quickly, if you need to pay someone to set it up.
Way to undermine an interesting product launch through poorly chosen language:
> Let’s be frank the single‑player notebook has felt outdated for a while now. We’re open‑sourcing its successor. Jupyter belongs in the hall of great ideas — alongside “Hello, world.” and “View Source.”
If you're trying to reach out to the Python community this is not the way to do it. Completely unnecessary hostile language there! Have some respect.
My advice to Deepnote is to scrap this launch announcement (ideally with an apology) and try again. They've built something genuinely useful and interesting but it's going go get a lot less attention than it deserves if they introduce the open source version to the world like this.
I don't know much about this, but I understand Project Jupyter is Nonprofit. If I go to "jupyter.org" I see a tab "Community" and another "Governance". If I go to "deepnote.com" I see "Customers" and "Pricing".
Why would people want a standard to be controlled by a private company? I don't think the "Open-Sourcing" of it says enough. How does licensing work with formats or standards?
I’m not familiar with Deepnote, but I have quite a lot of experience with Jupyter, and if someone were to ask me if there are more modern alternatives I would immediately point them to marimo (https://marimo.io/). For me marimo is already a successor to Jupyter, it has replaced it entirely for me.
I've been a long-time Twitter user. I don't hate Elon, so when he bought it I was cautiously optimistic.
I deactivated last week. The platform is bad and getting worse. It's scammy and spammy. Everything is designed around garbage engagement, so that the X team can brag about how good the product is doing.
Nobody wants a damn web view. If I'm clicking off to a link, I may want to click to another app and back in and still be where it was... If it's in a webview that's gone as soon as I click out. Yes, you can open in Chrome or whatever, open in a browser, but that's a pain in the ass to do an the time. I hate web views, in all forms.
I find it so sad that Twitter still gets traffic at all. Even if we put aside the super shady content on this platform (free speech, lol), the app, either on the web or mobile, has a sub-par user experience.
I wish all the devs that I respect were using another platform.
Alternative to the reject and request rewrite approach, which may not work in the corporation environment. You schedule a really long video call with the offending person, with the agenda politely describing that for such a huge and extensive change, a collaborative meeting is required. You then notify your lead that new huge task has arrived which will take X hours from you, so if he wishes to re-prioritize tasks, he is welcome. And then if the meeting happen, you literally go line by line, demanding that author explain them to you. And if explanation or a meeting are refused, you can reject RP with a clear explanation why.
When I worked at RSA over a decade ago, we developed Bloom filter-based indexing to speed up querying on a proprietary database that was specialised for storing petabytes of network events and packet data. I implemented the core Bloom filter-based indexer based on MurmurHash2 functions and I was quite proud of the work I did back then. The resulting improvement in query performance looked impressive to our customers. I remember the querying speed went up from roughly 49,000 records per second to roughly 1,490,000 records per second, so nearly a 30-fold increase.
However, the performance gain is not surprising at all since Bloom filters allow the querying engine to skip large blocks of data with certainty when the blocks do not contain the target data. False negatives are impossible. False positives occur but the rate of false positives can be made very small with well-chosen parameters and trade-offs.
With 4 hash functions (k = 4), 10007 bits per bloom filter (m = 10007) and a new bloom filter for every 1000 records (n = 1000), we achieved a theoretical false-positive rate of only 1.18% ((1 - e(-k * n / m)) ^ k = 0.0118). In practice, over a period of 5 years, we found that the actual false positive rate varied between 1.13% and 1.29%.
The only downside of a false positive is that it makes the query engine read a data block unnecessarily to verify whether the target data is present. This affects performance but not correctness; much like how CPU branch misprediction affects performance but not correctness.
A 30-fold increase in querying speed with just 1.25 kB of overhead per data block of 1000 records (each block roughly 1 MB to 2 MB in size) was, in my view, an excellent trade-off. It made a lot of difference to the customer experience, turning what used to be a 2 minute wait for query results into a wait of just about 5 seconds, or in larger queries, reducing a 30 minute wait to about 1 minute.
First, the author is actually wrong. The date is not 5 years before you submit, but is 5 years before the form is received by the home office! So there are a few days of uncertainty, depending on how fast Royal Mail was with the physical documents.
Additionally, I did a request for my information from the home office prior to filling in my form. After all, you have the right to request the information they have on you that will be used to verify your form. Kafka would be proud.
Let me tell you, Home Office doesn't have a clue where you were 5 years ago. It had approximately 50% of my trips, and frequently only had only one leg of the journey. Plane, ferry, train, sailboat, ... it didn't matter. It seems like they have not been keeping the information very well.
I don’t see how being critical of this is a knee jerk response.
Thinking, like intelligence and many other words designating complex things, isn’t a simple topic. The word and concept developed in a world where it referred to human beings, and in a lesser sense, to animals.
To simply disregard that entire conceptual history and say, “well it’s doing a thing that looks like thinking, ergo it’s thinking” is the lazy move. What’s really needed is an analysis of what thinking actually means, as a word. Unfortunately everyone is loathe to argue about definitions, even when that is fundamentally what this is all about.
Until that conceptual clarification happens, you can expect endless messy debates with no real resolution.
“For every complex problem there is an answer that is clear, simple, and wrong.” - H. L. Mencken
I had no idea travel was this difficult for people who aren't EU citizens.
Wow, I'm almost annoyed on the authors behalf of how much hoops there are to jump through.
>To apply for British citizenship, you need to prove you were physically in the UK on your application date but five years ago. Not approximately five years, not that week—that exact day when you press "submit" on the form minus five years. Miss it by 24 hours and your application is reject after months of waiting, and you have to pay a hefty fee to re-apply.
That's a hilarious requirement. I wonder how that ended up in there.
> buy a sausage roll at Greggs
If that's the first thing he thinks of while transiting through a UK airport, he deserves a citizenship, no questions.
Since we're doing PSAs, isn't it also now just a completely broken platform on mobile for everyone who isn't logged in?
> Something went wrong, but don't fret - let's give it another shot.
This is all I've seen for literally years now. No real error, does not even say to login or install an app, just blames it on my privacy extensions (I don't actually have any) and offers a button to pointlessly try again. No big loss, but surprising! On the one hand, it's the only time big tech isn't engaged in obnoxious harassment, but it's also a conspicuously dumb oversight in the funnel
The author makes an error while explaining Jevons’ Paradox. They say “the cheaper and faster we got at producing coal, the more coal we ended up using”. But that’s not paradoxical at all. That’s basic supply and demand curve from economics 101. It’s exactly what you’d expect.
Whereas if you take the extraordinarily difficult step of opening Wikipedia (https://en.wikipedia.org/wiki/Jevons_paradox), you understand the real paradox. Jevons observed that when steam engines get more efficient and individually use less coal, we end up deploying many more steam engines in many new industries and use cases, increasing coal consumption overall.
This seems like a basic oversight, maybe this guy hasn’t heard of Wikipedia. On the other hand, I think he wants to push the narrative that AI is seeing enormous productivity gains. He does this by using price per token, which has fallen. Very similar to Jevons supposedly observing “coal production”. But the author is struggling to point to new industries and use cases that were opened up by AI, like when we deployed steam engines back in the 1860s. So he misstates Jevons paradox, removes the paradoxical part, and makes it seem like his thesis makes sense.
Jevons’ Paradox could still apply here! I’m not saying it doesn’t. But we just haven’t seen the examples quite yet. A good example would be an observed surge in demand custom built software as software engineers become more efficient. But lower token costs ain’t it.
Open source? Close it and ask them resubmit a smaller one and justify the complexity of things like a DSL if they wanted it included.
For work? Close it and remind them that their AI velocity doesn't save the company time if it takes me many hours (or even days depending on the complexity of the 9k lines) to review something intended to be merged into an important service. Ask them to resubmit a smaller one and justify the complexity of things like a DSL if they wanted it included. If my boss forces me to review it, then I do so and start quietly applying for new jobs where my job isn't to spend 10x (or 100x) more time reviewing code than my coworkers did "writing" it.
> How would you go about reviewing a PR like this?
Depends on the context. Is this from:
1. A colleague in your workplace. You go "Hey ____, That's kind of a big PR, I am not sure I can review this in a reasonable time frame can you split it up to more manageable pieces? PS: Do we really need a DSL for this?"
2. A new contributor to your open source project. You go "Hey ____, Thanks for your interest in helping us develop X. Unfortunately we don't have the resources to go over such a large PR. If you are still interested in helping please consider taking a swing at one of our existing issues that can be found here."
3. A contributor you already know. You go "Hey I can't review this ___, its just too long. Can we break it up to smaller parts?"
Regardless of the situation be honest, and point out you just can't review that long a PR.
The whole situation is so extreme it almost reads like sick parody. Last year there were riots in Israel when some IDF soldiers were arrested for raping prisoners. The riots were in defense of the rapists, and were attended not only by extremist Israeli civilians but also Israeli lawmakers, who stormed the military base where the rapists were being held.
The leaker releases a video of some of the abuse and is then accused of "blood libel" against the IDF by the Minister of Defense, Israel Katz. That phrase, "blood libel", is specifically intended to invoke the old medieval stories of Jewish people sacrificing and eating gentile children for their religious holidays. For leaking a video proving that the abuse is real.
> I look forward to the "personal computing" period, with small models distributed everywhere...
One could argue that this period was just a brief fluke. Personal computers really took off only in the 1990s, web 2.0 happened in the mid-2000s. Now, for the average person, 95%+ of screen time boils down to using the computer as a dumb terminal to access centralized services "in the cloud".
I almost posted something this morning about this, because I received an email that really frustrated me. I have a 401k with Guideline from an old employer. The email was from Accrue , and said, in part:
> Login: Please visit my.accrue401k.com to log in. You’ll find that the 401(k) dashboard and user experience remain familiar. If you’ve set up your account, your same login credentials will provide you access into the dashboard. (Please note, Accrue does not currently offer a mobile app).
The my.accrue401k.com part was a hyperlink to that site. I've independently done some digging (and contacted my old employer to verify!) but this is precisely how a targeted phishing attack would work. Asking someone to enter their financial account credentials into a site they've never used or heard of, based entirely on an unsolicited email, is INSANE.
This email was the first time I've heard of Gusto, of Guideline being acquired, or of Accrue 401k (which apparently is the company created to hold Guideline's 401k accounts that are NOT affiliated with Gusto). Nice.
Not sure the dial-up analogy fits, instead I tend to think we are in the mainframe period of AI, with large centralised computing models that are so big and expensive to host, only a few corporations can afford to do so. We rent a computing timeshare from them (tokens = punch cards).
I look forward to the "personal computing" period, with small models distributed everywhere...
Having seen LLMs so many times produce coherent, sensible and valid chains of reasoning to diagnose issues and bugs in software I work on, I am at this point in absolutely no doubt that they are thinking.
Consciousness or self awareness is of course a different question, and ones whose answer seems less clear right now.
Knee jerk dismissing the evidence in front of your eyes because you find it unbelievable that we can achieve true reasoning via scaled matrix multiplication is understandable, but also betrays a lack of imagination and flexibility of thought. The world is full of bizarre wonders and this is just one more to add to the list.
> So, eventually & reluctantly, I have changed my mind: there will be another major version of htmx.
> However, in order to keep my word that there will not be a htmx 3.0, the next release will instead be htmx 4.0.
technically correct.. the best kind of correct
From the headline, one might assume he directly edited or locked the page when he just commented on the article's discussion page that it should have a more neutral tone.
"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim." - Edsger Dijkstra
The debate around whether or not transformer-architecture-based AIs can "think" or not is so exhausting and I'm over it.
What's much more interesting is the question of "If what LLMs do today isn't actual thinking, what is something that only an actually thinking entity can do that LLMs can't?". Otherwise we go in endless circles about language and meaning of words instead of discussing practical, demonstrable capabilities.
OpenAI is generating $13B a year in revenue. Let’s be generous and say $20B. They’ve signed commitments to spend something like $1.4 trillion on compute. An asset that to date has proven to have a hyper-depreciation cycle.
Someone has to come up with $1.4 trillion in actual cash, fast, or this whole thing comes crashing down. Why? At the end of all this circular financing and deals are folks that actually want real cash (eg electricity utilities that aren’t going to accept OpenAI shares for payment).
If the above doesn’t freak you about a bit at how bonkers this whole thing has become then you need a reality check. “Selling ads” on ChatGPT ain’t gonna close that hole.
I'm really torn -- you and your engineers should be excited to work on your codebase. You should enter it and be like "yes, I've made good choices and this is a codebase I appreciate, and it has promise." If you have a set of storylines that make this migration appropriate, and its still early in the company that you can even do this in 3 days, then by all means, do it! And good luck. It'll never be cheaper to do it, and you are going to be "wearing" it for your company's lifetime.
But a part of me is reading this and thinking "friend... if PostHog was able to do what they're doing on the stack you're abandoning, do you think that stack is actually going to limit your scalability in any way that matters?" Like, you have the counterexample right there! Other companies are making the "technically worse" choice but making it work.
I love coding and I recognize that human beings are made of narratives, but this feels like 3 days you could have spent on customer needs or feature dev or marketing, and instead you rolled around in the code mud for a bit. It's fine to do that every now and then, and if this was a more radical jump (e.g. a BEAM language like Elixir or Gleam, or hell, even Golang, which has that preemptive scheduler + fast compiles/binary deploys + designed around a type system...) than I'd buy it more. And I'm not in your shoes so it's easy to armchair quarterback. But it smells a bit like getting in your head on technical narratives that are more fun to apply your creativity to, instead of the ones your company really needs.
IIRC this is a misconception and bullies bully strategically to climb the social ladder and benefit from it.
It's possible that they do it because they learned pathological systems of behaviour from pathological family/social experiences, but even if fighting back against them is also shitty, it beats enabling them to keep doing it (especially to you)
You can't have a sustainable ecology when China sends a city of trawlers to devastate fish stocks around the world and then sends it all back to China. It leaves the countries that were depending on that fish suffering with no repercussions for China. As far as I'm concerned, those cities of fishing boats should be sunk because it's an act of war by China.
Completely absent is quantity of jobs. If e.g. ml engineer job postings go +40% from 200 to 280 and writer job postings go -50% (over two years) from 20000 to 10000, then we have a better idea of the impact.
Without those data this report isn’t really quantifying impact on “180M jobs.”
People consider google as a trusted partner whereas it is designed as a retail factory. Mass serving of millions and protectioms whose false positives can destroy the lives of thousands. Still they are statistically correct. Nuking everything instead of the offending service? Convenient fir them. Unavailable support reps? Convenient for them? Meaningless automated answers? Convenient for them. Its not a solid system that has defects, it was designed that way. Their unavailability and abrupt cruelty does not serve as cost optimisation, it serves as liability optimisation.
I would love to like Nextcloud, it's pretty great that it does exist. Just that makes it better than... well everything else I haven't found.
What frustrates me is that it looks like it works, but once in a while it breaks in a way that is pretty much irreparable (or at least not in a practical way).
I want to run an iOS/Android app that backs up images on my server. I tried the iOS app and when it works, it's cool. It's just that once in a while I get errors like "locked webdav" files and it never seems to recover, or sometimes it just stops synchronising and the only way to recover seems to be to restart the sync from zero. It will gladly upload 80GB of pictures "for nothing", discarding each one when it arrives on the server because it already exists (or so it seems, maybe it just overwrites everything).
The thing is that I want my family to use the app, so I can't access their phone for multiple hours every 2 weeks; it has to work reliably.
If it was just for backing up my photos... well I don't need Nextcloud for that.
Again, alternatives just don't seem to exist, where I can install an app on my parent's iOS and have it synchronise their photo gallery in the background. Except I guess iCloud, that is.
> Nobody’s actually run this in production
We do at Discourse, in thousands of databases, and it's leveraged in most of the billions of page views we serve.
> Pre- vs. Post-Filtering (or: why you need to become a query planner expert)
This was fixed in version 0.8.0 via Iterative Scans (https://github.com/pgvector/pgvector?tab=readme-ov-file#iter...)
> Just use a real vector database
If you are running a single service that may be an easier sell, but it's not a silver bullet.
"Microsoft shipped a broken “Update and shut down” toggle with Windows 10, and it never acknowledged it until now."
I guess they know what's best for the user base, and this was obviously deemed not important. But boy did they get Copilot integrated in everything post haste.
Typical Microsoft hubris.
Everything but the Metric System: “The new YASA axial flux motor weighs just 28 pounds, or about the same as a small dog.” :)
Could lead to significant efficiency gains for EV's, because 1/4 of the motor weight means better power-to-weight ratio... a lot of things will automatically get better.
YASA was founded in 2009, a spin out from Oxford University following the PhD of founder and still CTO, Dr Tim Woolmer.
"Over the decades that followed both of these technologies were explored. But despite the potential for weight reduction, smaller size, shorter axle length and increased torque, it was the difficulty in manufacturing the axial flux technology that limited its commercial viability, because the motor could not be made by stacking laminations, as with radial machines."
"The breakthrough innovation came by segmenting the axial flux motor in discrete "pole-pieces", so the motor could be manufactured using Soft Magnetic Composite material.
SMC can be pressed at low cost into a wide variety of 3D shapes. This removed the need for the complex laminations, overcoming the major manufacturing challenge of the axial flux machine."
"In 2025, after a £12m investment, YASA opened the UK's first axial-flux super factory, in Oxfordshire.
The opening of this facility boosts YASA’s manufacturing capacity, setting new benchmarks in e-motor technology and quality, and enabling production to scale beyond 25,000 units per year."
This is awesome. Lighter motors also make electric flight more viable
When I was around 15, I used to hang out with a guy who was much senior to me, and he would bully us sometimes. One day, when we were bantering, I cracked a joke that a third guy with us (who was my age) found funny and crackled. The bully grabbed my neck and choked me till I lost consciousness. I remember having memory flashbacks related to missing a train, and someone waiting at the wagon door, waving at me to hurry and jump in before it is too late. I remember feeling stressed about missing the train. The next thing I remember is slowly regaining consciousness to see the bully and the 3rd guy splashing water at my face, looking very amused.
I implemented the same behavior in a different Google product.
I remember the PM working on this feature showing us their research on how iPhones rendered bars across different versions.
They had different spectrum ranges, one for each of maybe the last 3 iPhone versions at the time. And overlayed were lines that indicated the "breakpoints" where iPhones would show more bars.
And you could clearly see that on every release, iPhones were shifting the all the breakpoints more and more into the left, rendering more bars with less signal strength.
We tried to implement something that matched the most recent iPhone version.
Love all of these tips. I've hosted dozens of events since moving to NYC and figured I'd add 5 more:
1. If this is a dinner party (or people are all seated), force people to get up and move in a way that they'll meet new people. Do this when you're about 2/3 of the way through the party. Some will complain - do it anyway.
2. Plan 1 (ideally 2) interludes. It can be a small speech, moving people around, changing locations, having people vote on something, etc. For whatever reason, they make the night more memorable.
3. Do your best to make introductions natural and low-pressure. Saying things like "you two would really get along" can put pressure on people - especially shy ones. Bring up something they have in common and let them chat while you back away.
4. Go easy on folks who cancel last minute. They often don't feel good about doing it and you don't want to add more stress to them or yourself.
5. More music != more fun. Some music is good, but if people can't hear each other, turn it down.
If you're interested reading more about this stuff, read The Art of Gathering by Priya Parker.
I feel like this is really an American culture thing where parties or dinner parties are mostly the responsibility of the host. In movies or TV there’s even a common theme of guests judging the host’s hosting abilities.
In Brazil you throw a party to people you like and they all have a hand in helping you, sharing the load. Everyone will be responsible for some part of it, all of it is organized informally, there are no real formalities to the event. No one cares about making a science out of it.
I’ve never heard of a person complaining about party quality or comparing hosting abilities.
I had a halloween party once. It had everything: Substance abuse, debauchery, a proper PA system shaking the floor, good lighting, costumes, good people we knew, random other people we'd never met before, a keg of beer, epic bean dip... everything but the police, somehow.
Anyway, my boss showed up. I don't know if I invited him or if he just decided to be there on his own. He was having a great time with everything, and then he went into the back where some folks were enjoying the not-booze.
It was at this point that I lost track of him.
His jacket was still there. His motorcycle was still parked on the front sidewalk. But he was nowhere to be found, and his phone went straight to voicemail. It was like he'd simply vanished.
"Fuck," I thought to myself. "I've only had this job for a few months."
It turns out that he'd walked home, a couple of miles away. He woke up the next morning sitting at the picnic table in his back yard, shirtless, in the rain.
After that, I always made sure that I invited him to my other parties -- and he always made sure to decline, and tell me that he was never doing anything like that ever again. I consider this to be a win.
I sure miss the kind of parties where they have to get an emergency court order to cut power the building at 3am.
I learned everything I need to know about throwing parties from Dave Barry.
If you throw a party, the worst thing that you can do would be to throw the kind of party where your guests wake up today, and call you to say they had a nice time. Now you'll be be expected to throw another party next year.
What you should do is throw the kind of party where your guest wake up several days from now and call their lawyers to find out if they've been indicted for anything. You want your guests to be so anxious to avoid a recurrence of your party that they immediately start planning parties of their own, a year in advance, just to prevent you from having another one.
If your party is successful, the police will knock on your door, unless your party is very successful in which case they will lob tear gas through your living room window. As host, your job is to make sure that they don't arrest anybody. Or if they're dead set on arresting someone, your job is to make sure it isn't you.
As someone who finally recently escaped bluetooth firmware development: yes, Bluetooth is leaking secrets and it doesn't even require any silly RF shenanigans. Almost nothing actually implements LESC. Apple refuses to implement OOB pairing, so no peripherals can force you to use it, so everything is subject to MITM attacks. The entire ecosystem is a mess of consultants and underpaid devs copy-pasting Nordic sample code, with no time or financial incentive to do more than the bare minumum. Never trust any product that moves sensitive data through Bluetooth.
Tom Scott on why Swiss Cheese was losing its holes (spoiler: the process became too sanitised - reintroducing impurities solved the problem).
https://www.youtube.com/watch?v=evV05QeSjAw
I'd guess that Heinlein was aware of it and scaled it up in his imagination.
The Roads must roll — they are the arteries of the nation. When they stop, everything stops. Factories idle, food rots, men starve. The nation cannot live without its Roads.
A thousand feet wide, level as a floor, strip after strip moving past in ordered procession. The slow strips on the outside moved at five miles an hour; the inner ones faster and faster, until the express strip in the center rushed past at a hundred miles an hour.
-- The Roads Must Roll, Astounding Science Fiction, June 1940.
The only thing windows has focused on has been dark patterns to force users towards cloud and figuring out more and more ways to collect data to sell ads.
I’m not naive, I know a ton of huge enterprises still run huge fleets of windows “servers” but I still find it hilarious that a supposedly serious server OS would default to showing you the weather and ads in the start menu.
I just made the switch. I had been developing on Windows for the last couple of years, mostly to get used to the ecosystem. I wanted to be able to write C and C++ like I do on Linux, without an IDE and with the native toolchain (i.e. no cygwin). On top of that, I play Overwatch every night.
Windows just seems to have zero focus on performance though. React based start menu with visible lag, file Explorer (buggily) parsing files to display metadata before listing them, mysterious memory leaks not reflected in task manager processes.
I installed Linux Mint. While it didn't just work (TM), and I had to go into recovery mode to install Nvidia drivers, it worked well enough. I can run Overwatch via Steam and pull comparable FPS to Windows (500 FPS on a 3090 with dips into the 400s). Memory usage is stable and at a very low baseline.
It is nice to come back to Linux, and with games I don't really have a need to run Windows anymore.
Git bisect was an extremely powerful tool when I worked in a big-ball-of-mud codebase that had no test coverage and terrible abstractions which made it impossible to write meaningful tests in the first place. In that codebase it was far easier to find a bug by finding the commit it was introduced in - simply because it was impossible to reason through the codebase otherwise.
In any high quality codebase I’ve worked in, git bisect has been totally unnecessary. It doesn’t matter which commit the bug was introduced in when it’s simple to test the components of your code in isolation and you have useful observability to instruct you on where to look and what use inputs to test with.
This has been my experience working on backend web services - YMMV wildly in different domains.
Generally the problem with carpark solar is the mounting solutions are low volume niche products that cost way more than traditional ground mounts. My biggest hope is that this policy creates a marketplace with actual competition that comes up with more cost efficient mounting solutions that make it practical worldwide.
Parking lots are horrible. They're butt ugly heat islands that take up way too much space. While adding solar on them doesn't solve the last issue, it does help mitigate the heat island effect and solar panels are no less ugly than asphalt. Plus it is power creation right next to where it is being consumed for minimal transfer losses. It's also much nicer for the vehicles parked there to be in the shade.
When I get my way reviewing a codebase, I make sure that as much state as possible is saved in a URL, sometimes (though rarely) down to the scroll position.
I genuinely don't understand why people don't get more upset over hitting refresh on a webpage and ending up in a significantly different place. It's mind-boggling and actually insulting as a user. Or grabbing a URL and sending to another person, only to find out it doesn't make sense.
Developing like this on small teams also tends, in my experience, to lead to better UX, because it makes you much more aware of how much state you're cramming into a view. I'll admit it makes development slower, but I'll take the hit most days.
I've seen some people in this thread comment on how having state in a URL is risky because it then becomes a sort of public API that limits you. While I agree this might be a problem in some scenarios, I think there are many others where that is not the case, as copied URLs tend to be short-lived (bookmarks and "browser history" are an exception), mostly used for refreshing a page (which will later be closed) or for sharing . In the remaining cases, you can always plug in some code to migrate from the old URL to the new URL when loading, which will actually solve the issue if you got there via browser history (won't fix for bookmarks though).
I don't understand the need to hammer in the point that Fil-C is only valuable for this tiny, teeny, irrelevant microscopic niche, while not even talking about what the niche is? To be clear, the niche is rebuilding your entire GNU/Linux userland with full memory safety and completely acceptable performance, tomorrow, without rewriting anything, right? Is this such a silly little idiosyncratic hobby?
The real lesson they should learn is to not rely on running images and then using "docker commit" to turn it into an image, but instead to use proper image building tools.
If you absolutely have to do it that way, be very deliberate about what you actually need. Don't run an SSH daemon, don't run cron, don't an SMTP daemon, don't run the suite of daemons that run on a typical Linux server. Only run precisely what you need to create the files that you need for a "docker commit".
Each service that you run can potentially generate log files, lock files, temp files, named pipes, unix sockets and other things you don't want in your image.
Taking a snapshot from a working, regular VM and using that as a docker image is one of the worst ways to built one.
I took an OS in college in 2006 and the big project that my prof required us to do was to make modification of visopsys. The software was primitive at that time but still had UI interface.
I emailed the author to ask some questions in my project. The author had connection with my prof and informed my prof about this. My prof told me that I was not allowed to ask the author regarding this project. So I had to figured out on my own.
It was fun to play around with and learnt how things work at deep OS level. It was a good memory for me :)
And you guys notice anything about my username? :)
I’ve seen many small businesses do well on TikTok and Instagram by eschewing all fancy graphics and technology, and just talking into their phone’s camera like a normal person. “Hey I’m Joe, I just opened a cafe down here. It’s always been my dream, etc.” The more quirky and human the video, the better it does.
I know this new tool looks to be for static graphics; but I do think the same thing applies. Not using AI-generated polished graphics will become a differentiator.
> Finally, we keep this file synced with an AGENTS.md file to maintain compatibility with other AI IDEs that our engineers might be using.
I researched this the other day, the recommended (by Anthropic) way to do this is to have a CLAUDE.md with a single line in it:
@AGENTS.md
I’m more optimistic about the possibility of beneficial AGI in general than most folks, I think, but something that caught me in the article was the recourse to mammalian sociality to (effectively) advocate for compassion as an emergent quality of intelligence.
A known phenomenon among sociologists is that, while people may be compassionate, when you collect them into a superorganism like a corporation, army, or nation, they will by and large behave and make decisions according to the moral and ideological landscape that superorganism finds itself in. Nobody rational would kill another person for no reason, but a soldier will bomb a village for the sake of their nation’s geostrategic position. Nobody would throw someone out of their home or deny another person lifesaving medicine, but as a bank officer or an insurance agent, they make a living doing these things and sleep untroubled at night. A CEO will lay off 30,000 people - an entire small city cast off into an uncaring market - with all the introspection of a Mongol chieftain subjugating a city (and probably less emotion). Humans may be compassionate, but employees, soldiers, and politicians are not, even though at a glance they’re made of the same stuff.
That’s all to say that to just wave generally in the direction of mammalian compassion and say “of course a superintelligence will be compassionate” is to abdicate our responsibility for raising our cognitive children in an environment that rewards the morals we want them to have, which is emphatically not what we’re currently doing for the collective intelligences we’ve already created.
I understand the pitch here ("it finds bugs! it's basically all upside because worst case there's no output anyways"), but I'm finding some of these agents to be ... uhhh... kind of agressive at trying to find the solution and end up missing the forest for the trees. And there's some "oh you should fix this" stuff which, while sometimes isn't _wrong_, is completely besides the point.
The end result being these robots doing bikeshedding. When paired with junior engineers looking at this output and deciding to act on it, it just generates busywork. Not helping that everyone and their dog wants to automatically run their agent against PRs now
I'm trying to use these to some extent when I find myself in a canonical situation that should work and am not getting the value everyone else seems to get in many cases. Very much "trying to explain a thing to a junior engineer taking more time than doing it myself" thing, except at least the junior is a person.
> Set transmit power to High
Do NOT do this if you live in a densely populated area (e.g. apartment complex). You'll create noise for yourself and everybody else. Classic prisoner's dilemma - a few people could be assholes and profit from it, but if everyone's an asshole everybody suffers.
General rule on TX power: start on low and increase only if you know (or can confirm) it helps. Go back down if it doesn't.
I think it's actually very simple... the paradox of choice.
You introduce somebody to your attractive single friend there's a real chance they hit it off and form a relationship. You introduce them to 100 attractive friends, one makes more money, one has a more stable job, one is flirtier, etc then it's both exhausting and none of them seem like a great opportunity.
I think there are certain basic psychological facts that are anti-standard-economics that nerds (and therefore tech companies) almost always tend to be completely blind to. This is one of them.
Using coding agents to track down the root cause of bugs like this works really well:
> Three out of three one-shot debugging hits with no help is extremely impressive. Importantly, there is no need to trust the LLM or review its output when its job is just saving me an hour or two by telling me where the bug is, for me to reason about it and fix it.
The approach described here could also be a good way for LLM-skeptics to start exploring how these tools can help them without feeling like they're cheating, ripping off the work of everyone who's code was used to train the model or taking away the most fun part of their job (writing code).
Have the coding agents do the work of digging around hunting down those frustratingly difficult bugs - don't have it write code on your behalf.
> Because nobody actually wants a "web app". People want food, love, sex or: solutions.
Okay but when I start my car I want to drive it, not fuck it.
Great news. Now maybe we can go on the offense for once. Work to enable constitutional protections against this sort of thing, and develop systems that can work around it if and when this comes back again.
There are places in the world today where only sneakernet communication has any semblance of privacy, so we need non-specialist tools that can provide privacy and secrecy regardless of local conditions. (I’d love to see more communication tools that don’t assume an always-on connection, or low latency, or other first world conditions.)
The question posed sounds like "why should we have deterministic behavior if we can have non-deterministic behavior instead?"
Am I wrong to think that the answer is obvious? I mean, who wants web apps to behave differently every time you interact with them?
> With increasing evidence that chronic exposure to PM2.5, a neurotoxin, not only damages lungs and hearts but is also associated with dementia, probably not.
PM2.5 is not a neurotoxin, that's an absurd thing to say.
It's literally any particles under a certain size. Whether it's a neurotoxin is necessarily going to depend on what the substance is made of.
Whether your PM2.5 exposure is coming from automobiles or wildfires or a factory, the potential outcomes may be different in different areas of the body. Heck, my PM2.5 meter skyrockets whenever I cook anything in a frying pan, because many of the aerosolized oil droplets are PM2.5.
There is a general problem with rewarding people for the volume of stuff they create, rather than the quality.
If you incentivize researchers to publish papers, individuals will find ways to game the system, meeting the minimum quality bar, while taking the least effort to create the most papers and thereby receive the greatest reward.
Similarly, if you reward content creators based on views, you will get view maximization behaviors. If you reward ad placement based on impressions, you will see gaming for impressions.
Bad metrics or bad rewards cause bad behavior.
We see this over and over because the reward issuers are designing systems to optimize for their upstream metrics.
Put differently, the online world is optimized for algorithms, not humans.