Started a new job, with Windows 11. notepad.exe now takes 3 to 4 seconds to load on my work system... (even after closing the last tab and reopening the program).
Hah, it even has in-app purchases, for AI writing...
I'm glad it all worked out for this individual. I hope more people live their lives like this as the dystopia progresses.
Unfortunately, especially in the US, exercising your rights, or even just reading every paper you're expected to put your name to, not only constantly pisses people off for some reason, but also puts you at a significant disadvantage compared to the people that never push back in the interest of not making waves, or even because "whatever it's fine."
The Broadcom business model (outside the chip business) had been pretty well known, and they don’t really hide it.
They are tech bottom feeders. They find large businesses with a decent moat and free cash flow but are in long term decline (and wasting cash trying to find something new). They buy them, cut development, support and marginal products. Raise prices and squeeze as much as they can.
I'm probably in the minority, but I do not want a "connection" with a business. I want transactional interactions that actually work.
That is something that AI is not giving us today. By design. Companies are not switching to AI customer service because it's better or cheaper for the same service. They are choosing to replace customer service with AI chat bots that simulate the customer service experience without actually providing the service part.
What is is about management where they can't see how bad and half-baked these customer service agents are, but the customer can tell INSTANTLY they're AI and just not helpful in the slightest
So this is the other side of banning American models for non-Americans? And how exactly do they plan on enforcing all of this? Great Firewall of America?
This is a complete joke. The malicious clowns behind this should be removed from power and prevented from ever holding any position of power in any form of governance system.
When I use a computer to do work I want the computer to be right. I want to be able to trust the computer. With the inherit non-determinism and probabilistic nature of generative-AI, that fundamental reason why I engage a computer is lost.
If the spreadsheet is wrong, it’s because the math is wrong, it’s because I made a mistake. It’s not because all of a sudden the computer decided the nature of algebra should be different than it is.
Part of the reason why humans are rejecting AI is that we are putting it in places where it makes no sense, or places where humans prefer a human in the loop, there are plenty of places where machine learning algorithms make sense, but customer service is not one of them.
The articles and discussion around this and the Amazon story all seem to me to be an earnest tech press and community searching for a genuine reason for the administration blocking Anthropic’s models.
However, thinking back to the spat with the DoD and more generally how the administration is much more supportive - and supported by - OpenAI and XAI and it’s easy to imagine this is just another escalation in the fight between a “liberal leaning” company and its competitors and the administration.
There might have been something said by someone at Amazon or something but I’d guess Occam’s razor the administration just leapt at the chance after their supplier sanctions fell flat?
As others have said, it's Apple and they do not take kindly to other people leaking their technology/announcements ahead of time.
See also: the time that ATI's CEO told his employees that their chips would be powering Apple's to-be-announced hardware a few days before the announcement. Steve Jobs responded by pulling all of ATI's hardware from its demo units at the announcement, not mentioning ATI at all, cancelling a joint demonstration of the Radeon card that was going to be in the system, and never partnering with ATI again.
I think a lot of medical diagnosis could be solved with mass data collection if it was cheap enough. Right now, blood draws are somewhat routinely done because they provide a lot of human-interpretable indicators from a small number of values, and there is some evidence that e.g. "dogs can smell cancer" etc. (i.e. some diseases cause detectable odors).
With a big enough data set of [all kinds of bio values, including ones considered irrelevant for that disease] labeled with diagnoses, I suspect we could get very fast and accurate automatic diagnoses, even from a limited data set currently considered uncorrelated. Rather than going to your primary care physician, you'd go into the standardized, mass-produced and thus reasonably cheap everything-scanner, and you could likely get a more accurate diagnosis (or at least "things to check") than the average doctor would be able to give you under the practical constraints they typically operate under (time, available information/diagnostics).
This goes in that direction, and I'm really excited to see where it goes. I could imagine that given enough training data, ML models will be able to pick up on minute details that make it possible to diagnose diseases that weren't historically considered ultrasound-diagnoseable from this kind of detailed ultrasound.
I think combining it with gas chromatography/mass spectrometry of e.g. breath or blood/sweat/urine samples would also have the potential to be a cost-effective diagnosis method - lots of data, probably not all too useful for human interpretation, but would open the potential to walk up to a machine, breathe into it, spit into it, pee into it, give it a swab, and have it come up with an accurate diagnosis without invasive testing. If mass produced, the cost of something like this could easily drop below the cost of a typical doctor's visit. (I googled it and it seems like GCMS is already used for some diagnoses, but screening only for a few specific diseases rather than "throw ML at it and try to diagnose everything").
If the robot appears to be bringing me a taco, it would probably penetrate all of my defenses. Grok is currently more likely than Claude to arrive with the taco without being stopped by an export control directive.
Up until 2019, Windows was my daily driver and had been for the prior ~20. years. I had been regularly ssh-ing into Linux machines, but it didn't seem like a place I could live. Then, in 2019, I built a PC and, wanting to get more proficient in Linux environments, I made it a dual boot setup with a Ubuntu desktop partition and a Windows partition, expecting I'd inevitably get frustrated on the Linux partition by sidequests debugging driver issues or setting up peripherals, unproductive yak-shaving stuff. I had to google a setting or two over the first couple days, but other than that, everything just worked on the Linux partition. Things opened quickly, things installed easily, and things I was worried about (e.g. nvidia and printer drivers) were either automatic or a one-time step so small I don't remember it. After a couple weeks, I noticed there hadn't been a single moment where I had to switch to the Windows partition, and a month after that I reformatted the Windows partition ssd and added the storage to the Linux partition.
If you have considered switching to Linux and worried that it would be a chore, give it a shot (if you have the freedom to choose). It has been polished and ready since at least 2019. I have to use a Windows machine for work and, like this New Outlook issue shows, MSFT has concluded most users can't or won't leave so there's no margin in improving UX and some margin in doing things that make UX much worse. I don't think I'll elect to have a personal Windows machine ever again in my life.
The institutions and trust that generations of Americans carefully built has been gleefully torched by cruel incompetents in the space of a handful years. The damage, physical and social, is incalculable. The unpunished crimes, endless.
The reconstruction, if it happens at all, will take decades. It was all so unnecessary, so foolish.
Let's Encrypt has been working normally for most of the day. There was a ~90 minute period during which some of our users would have received a higher error rate due to upstream networking issues, but the majority of requests were successful even during that period.
It seems our status.io notes are being misinterpreted as much more severe than they were intended to reflect.
Edit: Note that this was written in response to a previous submission title implying that Let's Encrypt was entirely down most of the day.
It's funny how you can look at obscenely overvalued Western AI companies, with capex's that are unsustainable, and along comes a system which pokes a gaping hole in this model, and the response is "Ya, but what do those sneaky Chinese have up their sleeve here?!"
Perhaps the question should instead be "Why are these Western AI companies getting insane valuations on dubious ROI, and how can these Chinese models run on a fraction of the infrastructure?"
>The only hope we have is that they're wrong, and their power to change things will be minimal.
What even is the optimistic outcome if they are right? What are we all working towards? Like do people think that these AI companies will create some superintelligence, suck up all the financial benefits of that, and then just decide to share it with the rest of us out of kindness? Because I legitimately can't see a realistic outcome that actually benefits society as a whole. It all ends with very few obscenely rich people getting even more obscenely rich. But I guess we could tell an AI to put ourselves in the new Marvel movie to pass the time since we no longer have any jobs.
Noteworthy that Z.ai, maker of the just released near-frontier GLM 5.2, has already been on the Entity List since Jan 2025[1]. Being on the Entity List does not mean all trade is forbidden. Broadly speaking it means American companies and individuals are not allowed sell them goods and services, but they are still allowed to buy from them and pay them.
AFAIK the Chinese AI companies barely depend on US goods and services, except for nVidia GPUs which were export restricted anyway, so it doesn't seem to be very consequential (see Z.ai). For the RAM maker CXMT it could be a lot more problematic though.
I have a little insight here from working with ophthalmologists.
When the article mentions "using a cancer drug to treat a leading cause of blindness" it's talking about using Bevacizumab (brand name Avastin) to treat macular degeneration.
Avastin and Lucentis are molecularly the same drug. The difference is that Avastin isn't packaged to be injected into eyeballs, and Lucentis is. Also, Avastin costs about $50/dose, and Lucentis costs around $1,500/dose.
There are compounding pharmacies that will repackage Avastin into a format that can be injected into eyeballs.
However, many ophthalmologists are hesitatnt to use Avastin because that process introduces the potential for contamination. Not that it happens frequently, but every ophthalmologist I know lives in fear of introducing an infection into their patient's eyeballs. And what level of risk is worth saving $1,450 per dose on an injection? 1/100? 1/1000?
I know at least one ophthalmologist who would use Avastin, but would culture a sample from every batch she got from the compounding pharmacy, and had at least one batch that cultured bacteria.
Assumedly, if she hadn't, every patient she injected with that batch would have been at risk of contracting endophthalmitis, and endophthalmitis is no joke. Not infrequently, it can mean losing vision.
I know several of that ophthalmologist's colleagues who had been using compounded Avastin to try and save their patients money, stopped after that incident.
I don't really have a conclusion beyond "drug pricing is insane and the issue is complex for providers."
> Me: "I don't like smartphone UIs. Everything is flat, nothing indicates where you can touch or not. I have to randomly try everything on the screen."
Response by non-tech person: "Well, yeah, of course you have to try everything? How else would this work?"
I think this goes deeper than many tech people realize.
From what I understood from talking with "nontechnical"(*) friends, relatives, etc, for a good potion of them, computers had always been "unpredictable magic". They got by through memorizing some very strict and rigid interaction sequences - "click this icon, then click that menu, then click that button, etc" and prayed nothing unexpected would happen. They were too scared and/or uninterested in computers to even try and find any rules or consistency in it.
I feel as if those nontechnical people "won" now. Now all UIs feel as inconsistent and unpredictable even for "techies" as any computer interaction felt to those people back then.
(* repeated from another thread: "nontechnical" in the "not fluent with PC use" sense, which is actually quite arrogant - they can have very high technical skill in other areas obviously)
As an engineer myself working for them 20 years ago, we were certainly not well paid like the article said. Quite the contrary: I was still on University(had not finished the final project) and had to do most of the hard technical work myself for someone else to just overview the results and sign. My salary was miserable.
Once I had finished I could earn 3 to 4 times more on several places.
They were also extremely creative taking foreign systems, studying the patent and modifying it to pay zero to the creators of the patents. This was done with things like the aluminium beams for electricity delivery that I think was developed by Italians, or the tunnelling machines that had all the pieces replicated inhouse.
Fun article, but it leaves out my favorite "almost ignore" feature in Git: `.gitattributes`.
This file lets you specify that git should "ignore" the diff from certain files. For instance, Node projects have a `package-lock.json` that is pure noise from a Git standpoint (it's just massive amounts of diff specifying specific versions of libraries, and the real human-readable version is in a separate `package.json` file).
With `.gitattributes` in the root of your project, you can just add a line:
`package-lock.json -diff`
Now, that file will still get staged/committed (which you want) ... but when you `git diff` you won't see the massive amounts of pointless diff in that file.
I'm surviving on consulting income for a wide variety of clients right now in this space, and let me tell you it's brutal and extremely difficult to get entry to this space for people that don't have a wide network and lots of industry experience. Academic experience typically doesn't count.
In addition there's a severe "passion tax" for these sorts of jobs, the salary difference for a "Data Scientist, Computational Biology", and "Computational Biologist" is pretty big, and hiring is also brutal.
I know a ton of extremely talented people who have been locked out of employment for a long time now. The high interest environment means that biotech investing has been hit extremely hard, as biotech is even higher risk than most software and AI spending (thanks for the correction, Schlagbohrer). Pharma companiees with big hits, like Lilly with GLP1 agonists, are hiring a bit as they try to move into the modern era of pharma with lots of AI tools, but it's still brutal.
I've always used ZFS because it's vastly superior to other options. When I see storage companies building without fault tolerance, or without a merkle tree (so that you can backup deltas efficiently without having to recompute them) it's a sign their marketing team has more influence over the company than their engineers.
Sadly, the few ZFS COTS options have been somewhat underpowered. QNAP supports ZFS filesystems, but their backup configuration won't let you arrange for a nas to pull from the source (instead of the source doing a push.) You can still pull it off by scheduling your own cron job, but this somewhat defeats the purpose of paying extra for a vendor solution.
UBNT is still supporting my 15 year old edgerouters with security updates, and their interface is clean and usable for anyone with basic network experience. And their video surveillance solutions are unusual in that they allow you to keep your footage entirely onsite and offline, an uncommon level of privacy. If they can bring the same polish to their storage solutions, I'll be using these new products for a long time.
There’s a very good reason we don’t test asymptomatic people in low incidence populations. Basically all positives are false positives when you do that, no matter how accurate the test is.
When you’re testing healthy randos for everything the odds of a positive being false have so many 9s it would make an SRE weep.
Unless this is accurate to a degree previously unheard of in medical science it’s a boondoggle, and I can’t help but notice there’s no mention of accuracy.
Hah, it even has in-app purchases, for AI writing...