Will somebody vouch for this? I know what you're thinking. This doesn't encourage curiosity! This is politics! Same old, same old.
HN! Come on, how myopic must you be to not see that a working democracy is a condition sine qua non of your dear curiosity. You see what's happening in Iran, right? Well, they don't!
One of the things I haven't seen discussed yet in the comments is this claim:
Clay-like design freedom
They're claiming this not only can fit custom geometries, but can be part of the structure itself. Would love to see what they're building this out of. I expect we'll see some people dissecting the verve cycle batteries soon enough.
AI makes coding plans that often come up with phases. It can be interesting to ask it to skip a phase, and do the next one. You can get interesting data about prospective other futures.
IMO using subagents to generate good context is a huge win. That doesn't really require a worktree. But once you have good starting places, good contexts you can feed into a LLM, there's IMO not much concern about "build it's own context" (it's already provided right here) nor "wastes tokens" (since is GoodPut / GoodTokens).
the workflow of how we feel and build contexts is the art of this all right now. this project is on point. it's going to be a boom year for Terminal Multiplexing.
I generally download from https://rutracker.org/ (need an account to search not for downloading). They have pretty much everything that you can imagine (not just films) and in proper quality too (BD Remuxes etc). There will be no scene releases here because they add russian/ukrainian dubs and subs to almost all films but that's a small problem.
The other one is Heartive which lists torrents from the DHT network with Magnet links https://heartiveloves.pages.dev/ You just click on the torrent icon in the middle top of the selected film and all the available releases will be listed in plain text. The only downside that you need to be familiar with the release tags
Last but not least https://nyaa.si/ if you have a slight interest in anything japanese from manga to anime to much more
The success so far is really just political, which has largely been shutting down debate and dismissing calls for some kind of cost analysis of what we risk losing in enforcing this.
Whenever someone brings up this stuff, the politicians take the tone that "we won't let anyone get in the way of protecting children", and this is in response to people who in good faith think this can be done better. Media oligopolist love it because it regulates big tech, so they've been happy to platform supporters of the policy as well.
Third spaces won't reappear because the planning system in most cities shuts anything down the moment someone files a compliant. They get regulated out of existence the moment police express concern young people might gather there. The planning system (which in NSW/Sydney is the worse) has only gotten worse since the 80s after the green bans. It was largely put in place to allow for community say in how cities are shape, which sounds nice but it's mostly old people with free time participating who don't value 3rd spaces, even if they might end up liking them. They just want to keep things the same and avoid parking from getting overly complicated (and this is a stone throw away from train stations and the CBD).
Third places can be fixed by reforming planning which is slowly gaining momentum via YIMBY movements, but this social media ban is just not a serious contribution to changing that. If anything Social media phenomenon like Pokemon GO contributed more to these third places lighting up.
Governance in Australia is very paternalistic, it's a more high functioning version of the UK in that sense. I think it might be in part due to the voting system being a winner takes all single seat electorate preferential voting system which has a median voter bias for least controversial candidates.
As a kid I always felt being in Australia you missed out on a lot of things people got to do in America, that has slowly changed as media and technology has become less bound by borders but looks like that being undone.
IBM have an absolutely stellar record of blowing acquisitions. The highly motivated newly acquired team will be in honeymoon phase for 3 months, and then it slowly dawns on them that they’ve joined an unbelievably rigid organization where things like customer satisfaction and great products don’t matter at all. Then they’ll be in shock and disbelief at the mind boggling Byzantine rules and internal systems they have to use, whose sole purpose is to make sure nobody does anything. Finally, the core IBM sales force will start to make demands on them and will short to ground any vestiges of energy, time, opportunity and motivation they might have left. The good team members will leave and join a former business partner, or decide to spend more time with the family. They’ll meet often at the beginning to relive the glory days of pre-acquisition and recount times where they went went above and beyond for that important early customer. But then these meetings will become fewer and fewer. Finally they’ll find a way of massaging their resumes to cast the last years as being “at the heart of AI infrastructure”.
I've played around with many self-hosted file manager apps. My first one was Ajaxplorer which then became Pydio. I really liked Pydio but didn't stick with it because it was too slow. I briefly played with Nextcloud but didn't stick with it either.
Eventually I ran into FileRun and loved it, even though it wasn't completely open source. FileRun is fast, worked on both desktop and mobile via browser nicely, and I never had an issue with it. It was free for personal use a few years ago, and unfortunately is not anymore. But it's worth the license if you have the money for it.
I tried setting up SeaFile but I had issues getting it working via a reverse proxy and gave up on it.
I like copyparty (https://github.com/9001/copyparty) - really dead simple to use and quick like FIleRun - but the web interface is not geared towards casual users. I also miss Filerun's "Request a file" feature which worked very nicely if you just wanted someone to upload a file to you and then be done.
OP deserves a lot of credit for having the courage to write this.
1. Software was supposed to be trans-industrial. If you knew how to write code, you could work anywhere in the industrial economy. This meant it would have all the same benefits (stability, industry-agnosticism) of management without the negatives (subjectivity, politics). Unfortunately, that didn't last. Managers took that from us and created a culture of oblique/inappropriate specialization because it's easier to do that than to admit that they don't understand what we do.
2. Our industry has become extremely anti-intellectual. There's a sharp phase change between what your professors groom you for (out of a legacy leftist hope, never realized, that if a leadership education is delivered down into society's middle; then the scorned middle classes will revolt against the elite) and the world of Work, which hasn't evolved in most places. Adam Smith called Britain "a nation of shopkeepers". Corporate America is a nation of social climbers. It's fucking revolting. The good news is that, after a few years, you get used to it and develop the social skills necessary to survive it.
3. I don't think the future is in the Bay Area or Manhattan. Those are great places to build your career and gain some credibility/savings/experience while society figures out where the future will be. However, if you want to build the future, California's not the place for that anymore. Forty years ago, Northern CA was where people went to escape the Mad Men nonsense. Now, houses in Palo Alto-- a suburb; we're not talking San Francisco-- are more expensive than many places in Manhattan. The future's going to come out of a location that's free from the high-rent nonsense that creates a work culture of subordination. The years that made Silicon Valley cheap were those in which few feared the boss because one could make living money doing odd jobs, the cost of living being so low. That's over now. The Valley is Manhattan (again, Mad Men) minus winter and with worse architecture.
4. Through all this, you gotta play the long game. Sure, you're not going to be able to do hard experimental mathematics. You may have to let that dream go. Just keep current/sharp enough to be eligible for interesting work when it comes up. That is doable. Things are terrible right now for cognitive 1-percenters who want meaningful, interesting work (i.e. an upper middle class salary isn't enough, and it's never "stable" for top-0.x-percent intellects because of the job security risks that level of talent implies) but they won't always be like this.
5. Relatedly, if you watch the social climbers, they don't do a lot of real work. If you get even passably good at their game, you can get by with a couple hours of focused effort and that leaves 5-6 for self-directed learning. (Don't write code that you'll use later on work time-- you don't own that-- but feel free to explore and just rewrite the code from scratch at home.) Don't feel wrong about doing this; it's a crooked game and that makes criminals of everyone. Work is (for 95+ percent of people) just about advancing your career; the other shit is stuff people say to distract the naive and clueless. That idealistic shit is a luxury of the extremely privileged, and you need to pretend it as a status signal, but don't believe your own lie. Proles have to take what they can. Just be smart about it. Stealing office supplies == stupid. It's illegal and wrong and dumb and you'll get fired. Making decisions that help your career but aren't optimal for others (who don't give a shit about you either) == smart, if you don't get caught. If you steal, make sure to take intangibles.
6. If you can, start getting up at 5:00 in the morning. Get some productive hours banged out before you go to work. If you can't go to bed early, then compensate by taking mid-afternoon naps in a place where your co-workers won't find you (almost no one gets anything done during those hours anyway). Relatedly, it's worth a lot of money to kill your commute. If you can't afford to live near work, then consider a different city.
7. There are jobs that aren't like the corporate hell being describe above. They exist, but they're not common, and they're probably extremely selective in the Bay Area. When you get one, hold on to it for as long as it's good and learn as much as you can.
Also requires the device allows backup of passkeys. The infamous post where keepass was threatened if they were to continue to allow users to backup their own keys.
This gave me an idea - what if we made a stable diffusion based AI that would replace unimportant faces (and possibly other identifying details) with different ones - I have seen that AI can do this and make the change unnoticeable.
That way people would be safe from having their personal likeness and whereabouts accidentally plastered over the internet (except when they want their photo to be taken), and the end result wouldn't look so obviously modified as blurring faces or licence plates.
They're idiots who are throwing good money after bad in order to keep from admitting that they made a mistake.
MAGA voted for revolution, and they got Hillary Clinton (Trump's political idol since the 90s.) No matter who you vote for, you somehow get Hillary Clinton.
Everything that he is doing is what he said that he'd stop, except deportations of illegal aliens, attacks on public sector workers, and closing slush funds disguised as foreign aid, which his base (and others) very much support. The cutting of the public sector turned out to be a scam (barely anyone was cut, they were cut stupidly) and the slush funds were just shifted directly to the Pentagon; and although he closed the border, he was barely deporting. The Epstein failure, the Ukraine failure, and the Israel failure are also painfully obvious, and obviously things he could end at will (except he's obviously implicated by Epstein.)
So what he chose to do was to escalate the public show, the extravaganza of deportations. He tried to do it in the noisiest, most publicity aware way by picking fights with the most dingbat Democrats, creating alligator swamp prisons, and unleashing military in the streets of Democratic cities.
Since his backers (and every politician's backers) in the arms and oil industries want him to overthrow Venezuela, he's tried to somehow link what is overwhelmingly popular (the deportation of illegal aliens) with something stupid and unpopular (regime change in Venezuela) by talking about "illegal alien gang drug terrorists."
Ábrego García was a mistake that he got caught on, that he stupidly made a poster child for this PR strategy, and he's too dumb to let it go.
And to be fair, he sort of can't because it makes you wonder how many other mistakes there are, and then you can't give him the benefit of the doubt to push through the fast deportations he's trying to do to get around deliberately obstructionist judges.
Even worse, García is an illegal alien, and most people who voted for Trump want him deported anyway, no matter how nice he is. But Trump committed to painting immigrants as dangerous when statistically they're less dangerous than American natives. It's the second generation that become criminals, because they catch up to the US rate for people in their economic strata.
He could go back on that dumb claim, but it's not only been his brand since the first election, it's now tied to his excuse for continuing the siege of Venezuela, and escalating it to an outright attack. Also, the conquest of Venezuela would be a real legacy, because the US has become a country of thieves, and Venezuela has a lot to steal.
This is the real consideration. He doesn't care about immigration at all, which is why he isn't deporting, and isn't going after employers (who aren't Korean) but making exceptions and extolling H1Bs to MAGA rage. Attacking Venezuela is just going to lead to more Venezuelans coming in (and is why they're here now), and they'll just get a "legal" process that lets in just as many as the illegal process, in order to kill wages. Whoever you vote for, you get Hillary Clinton.
I feel this article acutely. My mother has a house full of antiques, fine china, and silverware that she values enormously but has essentially zero market value. Most pieces wouldn’t cover my monthly electric bill.
Here’s my plan - you’re welcome to copy it:
1. Make a video documenting each piece and its story while she’s still alive. Get her to tell the family history, where items came from, what they meant to her. This preserves what actually matters.
2. Set aside exactly three pieces that genuinely speak to me. Not “might be useful someday” - just three things I actually want.
3. At the funeral, announce anyone can take anything they want to remember her by. Let family self-select what has meaning to them.
4. Donate the rest wholesale to charity. Tax deduction should be around $25k - likely more financial benefit than selling piece by piece, with infinitely less hassle.
This honors the emotional value without inheriting the burden. The video preserves family history better than storing unused objects. And it avoids the soul-crushing experience of discovering your inheritance is worth less than a tank of gas.
Having no expectation of privacy in public used to be a reasonable stance when there was a real time+money cost to extended surveillance, which meant that you still had a moderate amount of privacy unless someone was willing to personally target you and spend significant resources.
You either had to have a cop or a PI tail you, or spend time and effort talking to neighbors and acquaintances collecting information and correlating it, and it was much harder to do so secretly.
Technology has reduced the cost of surveillance by several orders of magnitude, and although the premise is unchanged - that you've never had privacy in public - the practical impact has changed in an extremely disturbing way.
I think we're long overdue to rethink and strengthen privacy protections in public in the US. Technological limits, and policy limits on specific implementations are better than nothing, but it's clear to me that surveillance will continue to get cheaper and thus your effective privacy in public will continue to erode until a culture and legal shift in public privacy expectations. I'm not optimistic about that.
For others, I'm sure parent knows: OKLCH is largely a bugfix for CILEAB. Both try to make a color space where even steps feel evenly spaced to a human. But CIELAB had procedural flaws in its creation.
See slide 19: https://www.w3.org/Graphics/Color/Workshop/slides/talk/lille... -- if you ask CIELAB to make "pure blue" (RGB 0 0 100%) become grayscale, the intermediate colors become purple to the human eye. The entire point of a perceptual color space is that that doesn't happen. OKLCH fixes that.
BTW, credit to Björn Ottosson, who basically side-projected a color space into the web standards and more: https://bottosson.github.io/posts/oklab/ ... folks like him are why we sometimes have nice things!
I have keratoconus, which is where a lack of strength in my corneas has resulted in their losing their proper shapes.
I have several focal points in each eye, randomly clustered together. And unfortunately, there is no correlation (or reason for a correlation) between my eyes.
Imagine not being cross-eyed with two focal points, but with well over a baker's dozen. Even if I could line up one pair of points between my eyes, any improvement would be indiscernible in the mess I see.
Because the focal points are clustered close together, their impact is less at a distance (it just feels hard/impossible to properly focus, like looking through very slightly warped glass), but it is devastating up close. For reading.
Without help, I see so many copies of all the letters, randomly and tightly stamped all over each other, I could stare at a short line of text all day and never figure out what it said.
And this after having better than 20-20 vision at all distances, for most of my life.
(Fun fact: if I am in a dark room, and look at one of those tiny power-on LED lights on some media equipment with enough distance that it is basically a point, I can clearly see all my focal points - and also a dimmer curvy, spaghetti crossover mess of focal Beziers between and around them. My corneas are neither convex or concave. They are chaotic. Evil.)
Fortunately, I have hard gas-permeable "scleral" contact lenses. They form a near perfect cornea for me, so when I wear them, my awesome vision and glyphs live once again. "Scleral" refers to the fact that they are wide enough to rest on the whites (sclera) of the eyes, to completely cover and fill out my lame natural corneas.
So I am in pretty good if inconvenient shape.
But I would absolutely love it if this new method allowed my corneas to be reshaped. Any improvement would be a big deal.
(There is surgery where corneas are soaked with a binder, which is fixed with a laser, that strengthens them and stops/slows Keratoconus from getting worse. But it cannot recover what has already been lost.)
This, along with the "CHALLENGE.md" and "ROADMAP.md" document, is an incredibly cool way to show off your project and to give people a playground to use to try it out. The game idea itself is pretty interesting too.
It would be awesome if I ... didn't have to deal with AWS to use it. I guess maybe that might be a good use case for agentic coding: "Hey, Kiro - can you make this thing just use a local database and my Anthropic API key?"
Complaining aside though, I think that's just such a cool framework for a demo. Nice idea.
> and would be a third core device a person would put on their desk after a MacBook Pro and an iPhone.
Or build an AI-enabled device that replaces both. All you really need is local sensors, local emitters, and lots of local+remote processing+storage.
The laptop/desktop mostly goes away, when most people won't need desks, since most desk-requiring jobs will soon be done passably by "AI". (Whether the "AI" is actual intelligence, or just robo-plagiarism of training material.) Do you really need a keyboard, when there's nothing for you to type. Do you really need a bunch of screens, when you're not looking at and reasoning about lots of information.
If anyone is going to build a one-device for the idle and disaffected eloi, to be harvested of remaining value, by the weathly, who increasingly consolidate all of the wealth and power, it may well be OpenAI building that device.
Apple isn't the best candidate to nail this, because they have lingering whiffs of hippie counterculture in their self-image. And for a long time, Google thought of themselves as the good ones, with vestiges of that enduring, no matter how much DoubleClick metastasizes. But OpenAI staff was confronted unambiguously with its true self early on, so doesn't have the encumbrances that the others do.
PostgreSQL uses heap files for the primary table storage, not B-trees. In PostgreSQL table data is primarily stored in heap files (unordered collections of pages/blocks). Indexes (including primary key indexes) use B-trees (specifically B+ trees). When you query a table via an index, the B-tree index points to locations in the heap file
InnoDB uses a clustered index approach. The primary key index is a B-tree. The actual table data is stored in the leaf nodes of this B-tree. Secondary indexes point to the primary key.
One is not better than the other in general terms. InnoDB's clustered B-tree approach shines when:
You frequently access data in primary key order
Your workload has many range scans on the primary key
You need predictable performance for primary key lookups
Your data naturally has a meaningful ordering that matches your access patterns
PostgreSQL's heap approach excels when:
You frequently update non-key columns (less page splits/reorganization)
You have many secondary indexes (they're smaller without primary keys)
Your access patterns vary widely and don't follow one particular field
You need faster table scans when indexes aren't applicable
I personally find PostgreSQL's approach more flexible for complex analytical workloads with unpredictable access patterns, while InnoDB's clustered approach feels more optimized for OLTP workloads with predictable key-based access patterns. The "better" system depends entirely on your specific workload, data characteristics, and access patterns.
It's about feeling satisfied and full too. Satiety. With the demonization of carbs and fat, protein is all that's left for calories. And its the most satiating/satiefying. Which every article like this seems to gloss over.
Protein does two things. The minimum amount required is necessary to replenish amino acids and proteins. The excess is spent on calories. At the end of the day, calories are sort of calories. The body needs energy and it needs one of the three and protein is really the third best of the three, but has the least bad reputation. This is where marketing has overtaken science and fact based decision making.
Complex carbs are the best energy but they need to be cut with bran and fiber to regulate their absorption speed. That kind of gets left out of the fiber discussion, that fiber is part of a pairing with carbs.
The seed oil hysteria (which is really only focused on omega6s/polys anyway,) along with the perception of "fat" making you fat, has steered people away from monounsaturated fats being a primary calorie source despite being cost effective, healthy, and quick to consume.
Another place where marketing/blognutrition has overtaken reality is the idea of every protein needing to be complete, vs just eating complete protein over the course of a week or day. Collagen is missing tryptophan, which is abundant in whole milk, yet collagen is wrongly extolled as "not a source of protein and shouldn't be counted."
The other part of satiety is learning mindfulness, and being ok with hunger, and being mindful of not letting hunger control behavior mindlessly..
3FS isn't particularly fast in mdbench, though. Maybe our FDB tuning skill is what to blame, or FUSE, I don't know, but it doesn't really matter.
The truly amazing part for me is combining NVMe SSD + RDMA + supports reading a huge batch of random offsets from a few already opened huge files efficiently. This is how you get your training boxes consuming 20~30GiB/s (and roughly 4 million IOPS).
802.11ad/ay on unlicensed 60ghz, our most economical option is to deploy Ubiquiti Wave Pros. We see real-world 2gbps+ speeds at 15km distances. We have Wave Pro, XG, and XR radios all throughout the network for multigig links, and 95% of our non-business installs are Wave LRs and Nanos. We can do up to 33gbps symmetric on 70ghz licensed bands on a single radio, and I have a number of 10gbps radios, but they're not cheap.
> You know, to just be grossly generalistic, you could put half of Trump's supporters into what I call the basket of deplorables. (Laughter/applause) Right? (Laughter/applause) They're racist, sexist, homophobic, xenophobic, Islamophobic – you name it. And unfortunately, there are people like that. And he has lifted them up. He has given voice to their websites that used to only have 11,000 people – now have 11 million. He tweets and retweets their offensive hateful mean-spirited rhetoric. Now, some of those folks – they are irredeemable, but thankfully, they are not America.
Hilary was only wrong about two things: It's not half of his supporters it's 100% and they're not "deplorables", they're monsters.
The 8541.* category exclusions are interesting. Does the US self-produce all required quantities of LEDs and piezoelectric crystals and doesn't need to import those? Is the exception on photovoltaic components to discourage American companies from producing solar panels?
Because all of their claims are pig shit. I am personally one to extend a large assumption of good faith. And I'm not some Democratic partisan but rather a libertarian who would say "both sides" as recently as 2020. But after years of steelmanning and trying to engage with where they are coming from, at this point the dynamic is irrefutable - claiming to want freedom of speech, claiming other American freedoms like gun rights, claiming to want law and order, claiming to care about spending or inflation, claiming to want small government, claiming to want accountability, draping themselves in the flag while claiming to love this country, claiming to be the victims of their fellow citizens' unwarranted criticism - all of these references to lofty ideals are just to abuse good people's assumption of good faith while running interference for autocratic authoritarianism, red in tooth and claw, as they burn this country to the ground with their whiny grievance politics.
Trump in I believe last night's interview with Fox (guy seems to do these daily, repeating the same lies and nonsense night after night, so it's tough to know if it's new or not) complained that every other country had "raped and pillaged" the United States.
Imagine being on top of the world, the richest large country on the planet, and declaring that really you're the victim. There is a fundamental blindness to realize that all of these things are not givens, and that things can get much, much, much worse.
And they are going to get much worse. And Trump's approval rating will probably dip to 45%. The US has committed to the bit of being a modern idiocracy that just like the excitement of "oh boy, what hijinx is he going to do today???"
I have tried a lot of local models. I have 656GB of them on my computer so I have experience with a diverse array of LLMs. Gemma has been nothing to write home about and has been disappointing every single time I have used it.
Models that are worth writing home about are;
EXAONE-3.5-7.8B-Instruct - It was excellent at taking podcast transcriptions and generating show notes and summaries.
Rocinante-12B-v2i - Fun for stories and D&D
Qwen2.5-Coder-14B-Instruct - Good for simple coding tasks
OpenThinker-7B - Good and fast reasoning
The Deepseek destills - Able to handle more complex task while still being fast
DeepHermes-3-Llama-3-8B - A really good vLLM
Medical-Llama3-v2 - Very interesting but be careful
"The years that pass eat up your margin for error until there is no margin left. The mistakes you make are no longer flaws of inexperience, they are flaws of character. To be young is to be constantly on the precipice of perfection – just a little further and you’ll get there – but you never get there, and suddenly you’re old, and find yourself in a permanent state of imperfection, which you must reckon with."
Because arbitrary, ignorant hack and slash "shock and awe"
is a smokescreen to conceal whatever it is they're really after.. privatization, emoluments, treason, and/or other arbitrary, political agendas.
Upstart's CLA definitely had an impact (Ubuntu even shipped an additional patch in the Ubuntu Upstart source package, because the author hadn't signed the CLA, so it couldn't be incorporated into the upstream), but Upstart had a pretty fundamental design flaw - the dependency model was entirely broken (https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/44765...), and Scott James Remnant left Canonical in early 2011 before rewriting it. Canonical never really funded anyone in a position to fix this kind of thing, which didn't really give a positive feeling about long term technical maintenance.
Will somebody vouch for this? I know what you're thinking. This doesn't encourage curiosity! This is politics! Same old, same old.
HN! Come on, how myopic must you be to not see that a working democracy is a condition sine qua non of your dear curiosity. You see what's happening in Iran, right? Well, they don't!
Put your glasses on, HN, you're not seeing shit.