I have never been interested in the “normal” social media apps like Instagram, Twitter, Tiktok, and the like. The content never appealed to me as a consumer enough to get started. Occasionally something would go viral enough that a friend would eventually link it to me and that was the whole experience.
Recently, I made a dumb little app for my kids and decided to try marketing it on social media just to see what it is like. It is fascinating in a sense and disheartening as well. I have been very unsuccessful, but the most signal tends to come from the dumbest content I have tried.
In doing this, I have come into contact with the social media feeds I never felt the need to look at and man… they are like a drug. I find myself mesmerized by random IG reels. It is one thing to understand what they are on an intellectual level and a totally different to feel it first hand.
Interesting. I haven’t fully read through the rule change, but seems like HHS is directly adopting the controls required by HITRUST? I have been out of the industry for a while. Always interesting how the industry shapes regulation and vice versa.
The best are the Jira tickets with a huge wall of AI slop requirements. Usually full of nonsense of course including implementation recommendations in the wrong language or framework. Questions for clarification met with blank stares from the author. Ah well, copy/paste into claude code and say “do this. make no mistakes” and get back to browsing HN…
I ran through the eval loop for a side project’s task (personalization of a micro video game, no thinking) last night. Head to head with Gemini 3 Flash Preview, results came out at basically a wash on my rubric. The output quality was good, well grounded, and reliable across 144 runs. But not noticeably better. It isn’t a traditional coding task, so can’t infer anything there. The amazing part was how fast it is. It was consistently about 2x faster than 3 Flash Preview and slightly faster than 3.1 Flash Lite Preview which is amazing. For my task, the price difference doesn’t matter, so easy upgrade. I plan to write up a quick blog post with the results over the weekend.
Data centers are such great targets in modern warfare. A few cheap drones can inflict billions in damage with low direct casualties (if the attacker even cares). I have heard AWS in particular is secretive about the exact location of their data centers, but no doubt every major country knows exactly where they are.
They are not widely publicized but they cannot be entirely secret because when you use AWS Direct Connect you need to know where to do the work in. Not every AWS datacenter has Direct Connect, but any OSS intel person could do the work required. When you have sufficient incentive to know where (for instance, trading crypto) you will find out the difference between, say, a matching engine running in AWS's datacenter run by AT Tokyo or by Colt.
AWS does not publicize all of their stuff, true, but they have partner documentation, their partners are proud to be vendors to Amazon and will advertise, and there are old unlisted PDFs on the Internet. Besides, Amazon has thousands of employees and contractors and their families. This kind of thing is within reach of any human being with sufficient interest.
Which is why peace and diplomacy is so important. The last thing we need is to be war hardening everything, which is likely impossible in this day and age.
Considering fasvists has track record of starting wars, failing at diplomacy or intentionally breaking it and then forcing everyone else to deal with the problem ... op claim and intolerance to nazi are consistent.
Some Paris data centers are disguised as apartment buildings with the classic Hausmannian facade, and then you open up Google maps and see a ton of AC units stacked on the roof. These aren’t likely major cloud data centers mind you, and the motivation for concealing them has more to do with the city’s aesthetic codes than military defense.
Some data centers are more valuable as targets than others. For example, those comprising us-gov-east-1 and us-gov-west-1 or, god forbid, us-east-1. I don’t expect it is a difficult task to find them and other critical infrastructure for a state, but probably more involved than popping open google maps.
That's a region. It's not only many buildings, it's many zones, each of which are many datacenters. A region is just a virtual partition for their services. A zone is a fault domain for their services, and a single zone is met by many datacenters, each of which can have many buildings. Or at the least, I know of at least one datacenter which has multiple buildings, that is within one zone that has multiple datacenters, that is within one region that has multiple zones.
Loudoun county in Virginia, near Washington-Dulles airport. Taxes on data centers built by AWS and other firms provide almost 40% of the county budget.
You can be secretive all you want, but it's extremely difficult to hide massive heat exchanging systems and/or generators from aerial/space photography. Particularly at the scale of an AWS-like datacenter.
Building a fully camouflaged datacenter could be done at much greater cost, but you still can't hide its thermal emissions from infrared. Basically every watt hour used in a datacenter environment turns into waste heat ultimately rejected into the atmosphere (except for the 0.000000001% that leaves the facility as photons down a fiber), so if you have N megawatts of waste heat from a rectangular shaped building located on a 300 x 400 meter sized plot of land, it's going to stand out.
Geothermal exists, but you would have to take care to design accordingly and even then there are plenty of other ways for a state actor to locate you. It probably doesn’t make much sense to spend money trying to hide from state actors; it’s probably better to (1) avoid conflict prone areas to the extent possible and (2) make it expensive for an attacker to shut you down (use more smaller data centers within a sensitive region, put some of them underground, etc) or (3) accept the risk of data center disruption.
Possibly but it is a bigger and more obvious project than just the data center itself, would cost an ass ton of money, and would mean the cooling line itself is a standalone target to take the data center down.
On the other side AMZN could have hired several RC plane hobbyists, fly them Emirates business class, put them into Burj Khalifa suites, fund several beefy jet, nitro or EDF planes for them (with jet getting as expensive as $5K), and have these guys on guard duty on the roof so they would take down any incoming drones (see Ukraine interceptor drones), and that would still be a pocket change compare to the datacenter damage. (of course somebody can get a startup going producing an automated container deployable unit consisting of like a 64 cell VLS with such interceptors plus radar plus optical - can be quickly deployed when necessary for example onto datacenters or say onto large ships navigating some treacherous waters )
More tech-y approach - AI (or even actual security guards) monitor the video cameras and once there are incoming drones, several MW of power can be redirected into those datacenter's large satellite dishes (more precisely - into very simple microwave generators installed on the dishes) and the dishes turned toward the incoming drones - the drones will get cooked in seconds, add the kitchen microwave sound effect.
> A few cheap drones can ...
It is temporary. The race is only starting. Soon you will have to have a hive of highly intelligent autonomous drones to have even slightest chance to make through a hive of highly intelligent autonomous interceptors, etc.
The government based defense departments are very slow and expensive though, while the extremely valuable targets like the datacenters belong to the transnationals and located across the world (and more and more in space). Thus the transnationals would have to take care of the defense of their assets themselves (or outsource it to other transnationals, like say imagine AWS providing air-defense-as-a-service), more efficiently and agile than the government defense departments. If you take a look at Palantir Karp's book "Technological Republic" you can read that between the lines there too.
They can typically operate indefinitely on diesel generators and have hot supply contracts with multiple suppliers. Even our small rinky dink datacenter had that.
So, it would generally be more effective to hit the actual datacenter than try to cut the power.
Several gulf state oil companies have declared force majeure on contracts they have to supply various customers due to the war. Good luck on getting diesel deliveries when things really hit the fan.
>> I wonder if you can uncover where the data center is just by using ping command.
Not exactly, but you can uncover cloud providers like Google and Azure, who forget to tell you, their "availability zones" are in the same data center ;-)
The relative lack of trucks is what would identify the data center. The only other buildings like that are warehouses, which have a lot more trucks going in and out relatively speaking.
That's some of how geolocation works. Ping can't go faster than the speed of light, so that gives you a circle for where something is. Ping from enough places and you can get a good enough idea, if you're the Iranian Guard or otherwise.
Maybe. I used to use this about 20 years ago, trying to track down where the UK root servers were so I could increase my drop-catching. Get test accounts on as many hosting providers as possible, check hops and ping times, move on to the next one. It's not as accurate as you'd hope, though.
Modern warfare mostly targets military weapons: missiles against missiles… or financial weapons: oil price steering. Lawyers pens against lawyer pens. even military casualties are avoided. There is 0 appetite for senseless destruction of civilian infrastructure. You are thinking ww1,ww2 which are not modern anymore.
Big tech's love for cheap labor is a great mechanism for finding where all their most valuable assets are and mapping out any and all vulnerabilities. I imagine state actors are applying to any and all low paying jobs that have seemingly juicy job requirements and feeling out details during interviews. Even better if you offer to accept a salary far below standard rates and actually get the job.
While probably not a state agent, I've personally done online interviews with some people that were clearly lying about everything and trying to feel out details about the company. People claiming to live in our country and being citizens but having little ability with the language, saying they would love to come to our city but it's a bit far, saying they graduated from a major university but being unable to describe anything about the town (with their resume mentioning graduating from a different university, and their LinkedIn a different university from either), random people moving around and arguing in the background, all their work was with random crypto businesses that shut down within months. I had to stop my coworkers from saying too much. I had to convince them why hiring that person for remote work and giving them access to our servers was a bad idea. There are without a doubt companies giving similar people physical access to their hardware. And there are undoubtedly people who practice interviewing to better deceive companies.
Opus 4.7 via code has been inconsistent for me. Sometimes, it feels like working with a brilliant collaborator and is as good as 4.5 and 4.6 were. Other times, it takes dumb and lazy short cuts. It can be quite frustrating. Its response when I tell it it did something wrong is often to write a memory... which is then does not always read. The inconsistency isn't due to session length or age either. These are all new sessions. I feel like sometimes, I get routed do a dumber model or some other hidden setting is applied.
My experience as well. This is even worse than just having a mediocre model, because I can work around that. The inconsistency means it produces different outputs for the same prompt, and I can't rely on that as a business tool.
It is scary building on the public cloud as a solo dev or small team. No real safety net, possibly unbounded costs, etc. A large portion of each personal project I do is spent thinking about how to prevent unexpected costs, detect and limit them, and react to them. I used to just chuck everything onto a droplet or VPS, but a lot of the projects I am doing lately need services from Google or AWS. I tend to prefer GCP at this point because at least I can programmatically disconnect the billing account when they get around to tripping the alert.
There are very few countries where consumer rights apply to B2B transactions, especially if it’s multiple people operating as a “small team”.
A solo dev however might be able to present themselves as a retail consumer, and leverage some trading standards related rules for unclear pricing or something similar.
Very interesting. I just started researching this topic yesterday to build something for adjacent use cases (sandboxing LLM authored programs). My initial prototype is using a wasm based sandbox, but I want something more robust and flexible.
Some of my use cases are very latency sensitive. What sort of overhead are you seeing?
Wasm sandboxes are fast for pure compute but get painful the moment LLM code needs filesystem access or subprocess spawning. And it will, constantly. Containers with seccomp filters give you near-native speed and way broader syscall support — overhead is basically startup time (~2s cold, sub-second warm). For anything IO-heavy it's not even close. We're doing throwaway containers at https://cyqle.in if anyone's curious.
So much content is just straight copy/pasted from the LLM now. Articles, blog posts, linked in posts, reddit comments, etc. Even just using the LLM for 'editing' tends to shift the voice to an obvious LLM voice when used naively. It is getting worse too. Last week a co-worker sent me a screenshot of Claude for me to review their "work", which was just whatever Claude made up.
Usually, if something is very obviously unfiltered LLM output, I just stop reading.
I do use LLMs for writing myself. They are useful, but are poor authors.
Recently, I made a dumb little app for my kids and decided to try marketing it on social media just to see what it is like. It is fascinating in a sense and disheartening as well. I have been very unsuccessful, but the most signal tends to come from the dumbest content I have tried.
In doing this, I have come into contact with the social media feeds I never felt the need to look at and man… they are like a drug. I find myself mesmerized by random IG reels. It is one thing to understand what they are on an intellectual level and a totally different to feel it first hand.
I miss MySpace.
reply