Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror

Comment Re: LOL (Score 1) 50

Audio production is another reason for me to still use Windows. I don't game.

Linux also has enormous accessibility issues for someone vision impaired, which Windows just doesn't have.

I use Lightroom classic also, which really has no open source equivalent - even on Windows, let alone Linux. No one does a better job than Adobe for printing photos, IMO. For cataloging, developing and editing, there are FOSS apps that are practical. But they don't integrate well with each other, the way LrC does. I hate that it's subscription, and would jump to FOSS if there was something that could compete. I suppose one could use LrC with Wine under Linux. But the visual accessibility issues would still be a blocker for me.

Comment Re: Not surprising (Score 1) 50

That's one reason, but the other one they claim is for security, which is improved by features of the more recent CPUs, implemented as core isolation and VBS. That's not good enough to justify dropping support for older CPUs, though.

One of the big reasons for me to use Windows is compatibility, in particular drivers for old or exotic hardware. Using them is still possible in Windows 11 - if you disable those new security features, which you still can.

You also can't fit new CPUs in old sockets on motherboards that still have slots for legacy hardware, such as PCI. Buying a new computer is not even an option. One has to stick with Windows 10, or hack Win11 to work on such machines, while it is still possible. Or give up on the old hardware entirely.

Comment Re:Not good at math (Score 1) 22

Unfortunately, they area lso a significant source of funding that schools and other state projects rely on, effectively a form of tax.

I will grant you that it is more fun to buy a ticket than directly write a check to the tax authorities voluntarily. It's also less efficient to administer.

Comment Re: How about a Linux distro (Score 1) 65

Of course, but there will need to be a ton of unsafe code in a kernel and drivers. It is difficult to write many of those things without using inline assembly. And even for the parts that you can, it may not perform well enough. If you have 25 to 50% of unsafe code in the kernel and drivers, the benefit from Rust is significantly reduced. But my main point is that you just aren't going to get a Linux kernel written 100% in Rust, let alone an entire distro.

Comment Re: Yoda's wisdom best again (Score 1) 172

I had to look up the reason for your mentioning FORTH. It's probably not the only programming language that has syntax not allowing the error-prone assignment and test in one go.

Yoda syntax is available in most common programming languages. I always use it regardless when writing code by hand. The reason is to ensure that security bugs are not introduced.

In theory, you can detect these problems with these compiler warnings, and code reviews.
In practice, not everyone turns the warnings on, or makes the compiler stop for all warnings. And code reviews are human and error-prone. With my vision decline now, I would like not catch an = vs == error during a code review. Vision is of the reasons I cannot work.

For compatibility reasons, you cannot disallow this syntax in existing languages such as C or C++. The Yoda notation makes a lot of sense to me. Able-people reviewers strongly dislike reading Yoda code, unfortunately. It is also against style code at some big tech companies that I worked at to use it. Fortunately, not at others.

Comment Re: Older students (Score 1) 35

You have to balance quality of life with quantity of life. You probably won' be able to get both.

The COVID pandemic probably helped the social security system a bunch, killing a bunch of retirees, and ending their benefits.
But many others have subpar health care, and will die young as a result, with their contributions paying for benefits for those in better health, that are typically wealthier. The thing is that Social security also covers disability. My SSDI benefit retroactive to 48, if eventually approved, would be higher than my retirement benefits at 62, since I'll likely never work again. Never mind that I started working in my teens, and have contributed the max to SS taxes for most of my working years. There won't be enough of those years. Yes, something has to give, and I believe it should first be for-profit insurance companies, phama execs, shareholders, and so on, before we start cutting benefits. My HIV meds cost $4000/month at Costco for example, without insurance. 600 euros for the same med in France without insurance. It would be very hard not to have insurance there, though. The only situation I could think of would be short-term traveling, and meds having been damaged/lost/stolen. Maintenance RX are not considered emergency, even if missing them turns out life threatening. So my husband couldn't get his HIV meds covered in Nevada when he forgot to take them on a trip, no matter how much he tried. It was a short-term trip, and he just had to go without.
For $4000 that the med costs in the US for an uninsured, you could also travel to Thailand for $700 roundtrip, buy the generic med for about $150, and fly right back, spending a total of $850, and saving the other $3150. But of course, it's illegal to come back to the US with those cheap generic HIV meds from Thailand ! Pharma really have thought of everything.

Comment Re: Older students (Score 2) 35

Yes, it's bad enough tat I stopped driving at night on my 48th, over a year ago. I can still manage it during the day. But commuting is out of the question, you have to be able to do it both ways. I haven't done any meaningful work post-COVID as I couldn't adjust to WFH. Even if companies all went full RTO, against the wishes of 95% of their employees, I would struggle to return due not being able to commute. There is no local transit or paratransit of any kind available to me. So, I could spend $100 a day on Uber or Lyft to commute - unless they decided to some day ban me, which they could do for any reason on a whim. You have to love unregulated American freedumb. But all this assumes that I could get another job in the first place, in this rotten environment with the so-called AI boom, which is very unlikely. I started coding for a living in my mid-teens, and had never planned to work for someone else past 50 anyway. I just didn't expect my career to effectively end in March 2020 at 43 with COVID/WFH - it was a surprise to find out that work environment was not suitable for me.

I survive now mainly because I invested very well, including my father's inheritance, extremely tax efficiently. Which is why I'm now legally a multi-millionaire on food stamps. The Social Security Administration still considers me to be able-bodied. SSA won't fast track SSDI unless you have a terminal condition, which I fortunately don't have. Someone not very smart recently decided to cut 10% of the staff at SSA, which means my disability case probably won't be heard by a judge for >2 years, and it will only succeed if I get a sympathetic one. Meanwhile, the "big beautiful bill" could cost me my healthcare, since "able-bodied" people , according to SSA, not your doctor, won't be eligible for Medicaid anymore. If the market doesn't crash again, I could sell enough stock to realize capital gains, and make sure to be on Obamacare and not Medicaid. Assuming the Supreme court allows the ACA to continue to exist. But I would lose the food stamps. You just gotta love the American system, choose between food or healthcare. Or housing, but my mansion is paid off.

Fortunately, I don't have anything else medically serious going on besides the MD and osteoporosis, like the 31 years with bipolar, or 19 years with HIV, so losing my healthcare would be no big deal, and not in any way life threatening. I was born and live in California, but hold two passports. My state has been under siege by a psychopath. I may end up self-deporting.

But coming back to Alzheimer's, there is really nothing that can reverse it. I don't know about memory care. But even the pharmaceuticals on the market don't pretend to do anything but help stop further damage. I could also tell you how well Preservision is working to stop my MD, too. At least it's not horribly expensive. Basically with Alzeimer's, I would probably want to have somebody send me to Bumfukistan and find a doctor willing to do it for the right price. It's really hard to make those arrangements ahead of time, though, even if there is a spouse willing and able to deal with it, which would be horrible even if possible. I hope it never gets to that. My grandma only lasted 15 years with Alzheimer's because my mom was a full time caregiver for her. She likely wouldn't have lasted 3 years in a nursing home. I have no idea how my mom is still alive after being a 15 year caregiver once my grandfather passed away, and my grandma's Alzheimer's was discovered. She might well have had it longer.

Comment Re: Older students (Score 1) 35

Oh absolutely, we aren't going to see 500 year old humans in either of our lifetimes. And possibly no future generation every will, we just won't know.

I just turned 49, and already have very old people issues - macular degeneration, and osteoporosis, both caused by prescription drugs, one of which arguably saved my life. These medication risks weren't known for one of them. They were known for the other.

I will live, but working again is probably not in the cards, and for other reasons as well. I don't know if I want another 49 years this way. I would be inclined to say yes today, still, if I could freeze my body in time, and nothing got worse from here on. That is just not possible. I may know when it's time. I do know I don't want to be a vegetable with Alzheimer, like my grandma who lived with it for 15 years until 94. Just unplug me, at that point. But most likely no doctor in any country ever will, under current laws. My dad made it all the way to 67. My mom is still kicking.

I don't think normalization will happen. Maybe in 250 years I could master and record all of Bach's keyboard works. So far I'm at about 5% by BWV numbers after 25 years. Sounds like I'll need all 500 years. There have been excellent blind musicians, and perhaps I would get better at music if I totally lost sight.

Comment Re:its just auto predict (Score 1) 235

I have no idea what those grades even mean, as I have never been to college or graduated high school.

The LLM rate of hallucination is extremely high. It's inherent to the design of LLMs. LLMs will never fully replace a single employee. You need someone to write the prompts. And you need someone to evaluate the output, and tell whether it is correct, outright wrong, our completely made-up but plausible bullshit. It takes skilled humans to be able to do that. The LLM will never achieve anything on its own. You cannot just give it even a single prompt, and then let it iterate on its own until it figures things out for you. It is not a full replacement for anyone.

What it is, in the right hands, is a tool for productivity improvement. It lets me prototype code very quickly, in ways that I could never do myself, in any amount of time, even in programming languages that are essentially foreign to me. Someone calling it autocomplete would be completely missing the point. It has other use cases, and again, in the right hands, will improve productivity. Whether that improvement is 5% or 95% remains to be seen. Even after a very successful >30 year programming career, I can say that it's much closer to the later for me than the former. I have never been accused of being a 10x programmer. I have always taken my sweet time. I can't think one of one deadline I have ever met. But the quality of what I produced was always sufficiently redeeming to keep at it. Now, I can do things 10x faster - and do things I just would not have attempted on my own without assistance. If you were already a 10x programmer, I don't know if the LLM would help you or hinder you. But it wouldn't make you a 100x programmer. And perhaps not even 10.1x.
The reality is I can tell the LLM about the program/code I am designing. It will write some stub code in a language I specify. It may or may not compile, until dozens of iterations. I can tell it which SDK to try to use to accomplish the task at hand. It will search public API definitions, SDK headers, etc. find the API names, generate function calls that may or may not be correct. But it still saves a lot time that I would otherwise spend doing research and reading myself just to find out the right call. I still do that research when LLMs can deliver only hallucinations, but not nearly as often as before. If in the end, the end code ends up a pile of non functional full of bugs crap beyond redemption, I can just throw the whole prototype away, having wasted only a few hours of my time wasted, and a bunch of AI CPU cycles. The LLM writes about 90% of the code. But it's always been the last 10% that take 90% of the time - to review, test, debug, etc. That still requires a human pilot. The LLM just can't do that alone. But it can give some insight based on program output, quickly add logging for you in suspected buggy areas, etc. Repetitive stuff that you could easily do by hand - if you know the language, but would take time. It doesn't suck at refactoring, either, another hugely tedious task when done fully manually. It is pretty satisfying when you end up with a working program source, but full of boilerplate copy-pasta from the LLM, and then can point out all the redundancy one by one, and then eliminate that copy-pasta, one bullet point at a time, in iterative fashion, until you cut 25-80% of the code, and the program is still equally functional, but also now much more maintainable.

The problem is of course that these productivity gains aren't match by corresponding demand increases. So, the end result is a reduction in overall labor force. Not its elimination, but a significant reduction.

Comment Re: Older students (Score 1) 35

Medical science already has provided enormous life expectancy, through vaccines, cure for some acute conditions, treatment for chronic conditions, and so on.

There is no reason to believe that will not continue, since research and science will continue to be fully funded, and these medical breakthroughs will be available and affordable to all who need it worldwide. /s

Comment Re: AI = glorified search engine (Score 1) 235

Infinitely less. I would never even get started with the project without the LLM. I don't know Python syntax well enough. I dislike the language, especially the indentation based logic, especially with my visual impairment now, but I hated it before that happened. Maybe I could learn "Python with braces" some day. But not regular Python. In any case, the only reason I used it is because so many APIs are readily available through Python libraries. It also would take me forever to RTFM.
Truly, I have written a ton of code I never could have dreamed of doing without the help of LLM. I'm not saying iterating over and over is fun. But the result speaks for itself.

That said, it's all been fairly custom personal projects, such as code to manage my very large home network with 350 IP devices, integration between pfSense and Unifi APs, command-line tools to migrate data between the too, operate on clients in bulk, in a way that the vendor web GUIs just can never do. When I polish it, I will put it on Github, as I think these tools may also be useful to others. Many other home users and small businesses combine the use of pfSense router and Unifi Wifi APs. I will not work on any features to support equipment I don't use, such as Unifi gateways, Unifi management switches, etc. And I will not offer support. But with open-source, maybe someone else can fork it if they need to.

The one I did last night with PowerShell monitors the Windows NIC traffic in real-time. And if it's above a set threshold, keeps my power hungry desktop PC from automatically going to sleep. This is so that my LAN backups can complete, the encrypted snapshots can upload to Google drive, etc, as these programs lack useful power management. Helps with traffic from my Torrents VM too ;) The PS script uses native Windows APIs to monitor the NIC stats, and keep the machine from sleeping. This is done through inline C# code making Win32 API calls.

The thing is, I don't know any PowerShell. I also don't know any C#. My main languages are C99, C++98, x86 / x64 assembly. But I was able to get all that stuff to work in languages basically foreign to me, due to excellent testing and debugging skills. The LLM certainly made tons of mistakes, like copying the wrong constants from the Win32 SDK and hardcoding them into the C# in PowerShell. That caused not only the machine to stay awake with network activity, but also the 3 power hungry monitors. All because a 0 was supposed to be 1. The LLM could never find that bug. But it was obvious to me when I read the code that copying & hard-coding stuff from an SDK was very error-prone, and I found the bug pretty much right away once I read it. It all comes from the fact that constants cannot be exported from the SDK. There is just no proper binding mechanism to do so ! No IDL, include file equivalents, etc. Only external API calls and structure definitions. But not constants. And wrong constants obviously produce bugs.

My next step was to ask the LLM which languages do have useful Win32 bindings, besides C, C++ and C#. The answer was Rust. So, I explored converting the working .ps1 script to .rs . So far, I still haven't succeeded in getting it to work, and I spent much more time on that than writing the original Powershell and getting it to work. Rust seems to be a major pain point for LLMs, so far.

I say LLMs, because I used both Gemini and ChatGPT to work on this, alternating between the 2 when I ran into query limits with one or the other. I would have switched to Github copilot if I hit both of their limits. Those are the only 3 LLMs I have tried for coding, so far. I know there are others. I will probably try them some day.

Anyway, the LLMs are extremely useful to me to implement these personal projects. Since I stopped working for a living, for various reasons, mainly COVID WFH changes I could not handle, psych issues, and vision issues, I have written a lot of code - but none collaboratively. Most of it will never become public, but some will. I have a hard time imagining how one would use the LLMs in a more collaborative context. The public ones are not going to respect a company's language style guide. They are going to swallow the code, train on it, and leak it to others, which is a big no-no for most commercial entities. Companies need to be be running their own private LLMs and train their local codebases. They would also need training on public internet data to be useful. That's probably what big tech is already doing. Small tech could likely not afford to. In any case, I don't think I will ever work for somebody else again. But the LLMs really do help get prototype some small stuff very quickly. They still hallucinate a very large % of the time, in coding or non-coding contexts. But the non-hallucinating output remains very useful.

Slashdot Top Deals

In a five year period we can get one superb programming language. Only we can't control when the five year period will begin.

Working...