On paper, Google should never have allowed the ChatGPT moment to happen ; how did a then non-profit create what was basically a better search engine than Google?
Google suffers from classic Innovator's Dilemma and need competition to refocus on what ought to be basic survival instincts. What is worse is the search users are not the customers. The customers of Google Search are the advertisers and they will always prioritise the needs of the customers and squander their moats as soon as the threat is gone.
Google allowed this to happen because they listened to their compliance department and were afraid of a backslash if LLM says something that could anger people.
Exactly, Google's business isn't search, it's ads. Is ChatGPT a more profitable system for delivering ads? That doesn't appear so, which means there's really no reason for Google to have created it first.
There’s a very negative immune response to the idea of Netflix running ads.
And yet they’re there, in the form of prominent product placement in all of their original series along with strategic placement in the frame to make sure they appear in cropped clips posted to social media and made into gifs.
Stranger Things alone has had 100-200 brands show up under the warm guise of nostalgia, with Coke alone putting up millions for all the less-than-subtle screen time their products get.
I’m certain AI providers will figure out how to slyly put the highest bidder into a certain proportion of output without necessarily acting out that scene in Wayne’s World.
I suspect google can last much longer in regards to an AI model chat engine that competes with open AI and other companies, without needing a profit from that particular product in a timely manner. I can's say the same for the others. Google is using it's own money to fund this without mch pressure for immediate profit in a time deadline. They can rely on their other services for revenue and profit for the meantime.
Google had an in-house chatbot that was never allowed to launch. I used to think that they were wrong but now I'm pretty sure they were right to not launch it. Users are very forgiving with a newcomer but not with an established company.
Think about it in terms of the research they put out into the ether though. The research grows into something viable, they sit back and watch the response and move when it makes sense.
It's like that old concept of saying something wrong in a forum on purpose to have everyone flame you for being wrong and needing to prove themselves better by each writing more elaborate answers.
Does anybody actually work with H100s and the like? Their failure rate is so high, I dont understand why anybody will even consider it feasible to put the machines in orbit or even the sea. By my ballpark estimate, if you have 800 H100s, after 6 months, about 100 would be overheating or throttling, and a few will disappear and one or two will crash the machine with load.
> Does anybody actually work with H100s and the like?
They don't. The expectation the cloud develops in people is that magic computers just appear. They're living at a virtualized layer where all the nitty gritty of real machines going down and needing to be serviced all the time is handled by unseen minions (sorry SREs and DC staff) and cluster management and provisioning software.
The reality is that datacenters in space is mind-boggling stupid, just from the infeasibility of maintenance alone.
During my army days, the sergeant major always seem to know where we would fail to clean during inspection standbys, eg the top rim of doors. Part of it is a hazing ritual, but it also means if you know where to look, you know where people will consistently fail.
As an SRE who previously had to manually inspect changes and releases, I quickly learn what to check for, and saved many production issues from happening, but I guess nobody will know about the failures that didnt happen, but they will notice the delay I introduced and the inspection process was automated together with the CD system and I am cut out. Fingers crossed the automation is as thorough or can learn common failure modes.
I've been in your position before, and it is indeed thankless. So, for what it's worth, thank you for all of the disasters you prevented. I believe you, and I appreciate it.
Oh, yeah, Obama being aloof was why the white men who questioned his citizenship openly - who are now entirely complicit in or supportive of an unaccountable gestapo randomly kidnapping people from the streets wth no ID or due process based on their skin tone - weren't "charmed" by him.
Even though I worked for companies that killed Sun, I never stopped admiring the foundational work the company was doing, which was not just cool, but critical for technological progress, and was very sad when the company sold out to Oracle and was gutted alive. HPC stuff Sun pioneered is still very relevant today. In an alternate timeline, Sun fully embraced Open Source and became a key pillar of the internet today.
Unfortunately, while we are well aware of cool tech companies that were ran aground by the finance/sales/management consulting types, Sun felt like a company ran aground by engineers.
Zuck famously kept the Sun logo up for quite a while when Facebook bought Sun's HQ campus, as a warning to the employees of what they could become. In some ways, Facebook/Meta is the spiritual successor of Sun, just like Google became the spiritual successor of SGI when they bought the SGI campus.
But these two ad driven companies never quite became the new Sun/SGI, for better and worse.
I don't disagree that Sun was a company run aground by engineers -- though I certainly like to think of myself as one of the engineers trying to navigate us around the rocky shoals! For whatever it's worth, I broadly stand by my analysis on HN fourteen years ago (!!) of Sun's demise[0] -- which now also stands as clear foreshadowing for Oxide eight years before its founding.[1]
There's a world where Sun did what you hoped (became a systems company) and created Joyent in-house. However, hyper-scaling means going fast and cheap before good comes along. Sun's habit was fast and good and that's an extremely difficult hurdle to overcome culturally. (By fast I mean growing a platform, not raw performance, FWIW).
Solaris 10/11, with all its technologies (zfs, zones, crossbow, dtrace, etc), was the pinnacle of UNIX that came out just when the world changed. At a company I worked at circa 2008-12 (that was a solaris shop) we essentially created a proto-docker with containers and ZFS that allowed rapid deployments and (re)building of our systems. It was a game changer for on-prem.
I'm not sure what Sun could realistically have done to come out the other side of the dot-com carnage. Other companies in roughly equivalent situations come to mind. You start looking at doing a hard reboot when the margins for that reboot aren't there and it's difficult to see the light at the end of the tunnel. Maybe an earlier reinvention involving more open source and alignment with where hardware was headed. Don't know.
Sun did waste a lot of money in buying MySQL, $800 million in cash and $200 million in stock. Certainly a distraction, as well.
Sun never offered any way to inexpensively get onto the on-ramp of Sun hardware and software as they thought they could continue selling high-margin hardware forever; they had their $995 V100 which even included their much-loved LOM which was a remote-management device like iLO/DRAC/IPMI , then followed it up with: nothing.
Some more companies undisclosed and of course in 2000 Cobalt Networks for $2.0B.
But in general, just hanging around on SPARC far to long. Unfortunately the person put in charger of SPARC told Scott that he thought SPARC could be saved but it would need 4-5 years. And that's when they went into mulit-core, selling everybody on the whole 'threw-put computing' nonsense.
Well, in stock market terms the MySQL deal paid for itself. It pushed the stock well up. However turning this in real money wasn't possible in the year they had till IBM and Oracle did their bidding.
That Solaris/Toshiba laptops deal was interesting, but if I recall correctly the price was a bit too much, and maybe it could have been considerd yet another distraction.
I surely would have liked to get one of those laptops, though.
I think the easiest thing would've been to basically ignore dotcom and thus only take a hit from the general financial downturn and not from basing your company around the stock bubble (tortoise vs hare type of deal) , but Platt is the only example I know of and he got kicked out of HP for doing that.
Honestly, it's because of what Sun's innovations in systems software that I look so fondly on their work.
I do ask myself after reading the HN comment you linked, how often is the limiting factor of systems software the hardware? Potentially a case of this with consumer hardware is ACPI issues, like [1] and [2]. You could design the best software, but if your underlying firmware or hardware is faulty, then you would have to design your software around the faults instead of improving the lower layers or accept bugs.
Oxide describes on their website issues with "vendors pointing fingers with no real accountability, even when teams need it most," and I have seen this point discussed online in regards to Oxide's work on designing their own hardware and firmware. Incidentally, I applied to Oxide recently; I think they're cool for the reasons I thought Sun was cool.
“ Believe me that some of us understood this: I worked extensively on both Solaris x86 and with the SPARC microprocessor teams -- and I never hesitated to tell anyone that was listening that our x86 boxes were starting to smoke the hell out of UltraSPARC.”
Was that before or after you realised the Linux kernel devs were better at squeezing performance efficiencies out of x86 than you guys were?
Why are you hurt? And why does it lead to comments such as above? I think you need to figure that out, because it wasn't a good wholesome comment by any measure.
I'm not. But there was a technical discussion made where the kernel devs at the time explained why they were beating the pants off Sun, and Cantrill replied with "have you kissed a girl".
That's the sort of behaviour I'd sack the guy for if he worked for me.
Oh brother, this again. To bystanders wondering what the hell this is about, it's actually about two things, the most recent of which was over a decade ago.
The first thing is a regrettable quip of mine on Usenet (RIP) from October 29, 1996 -- just over 29 years ago (!!) as of this writing. As I have made clear several (many?) times over the years, I do very much regret it: I was young and it was stupid.[0]
But that's also not what this is REALLY about, because that decades-old quip on Usenet had itself been forgotten for over a decade when it was dug up in 2013 by people who newly discovered that they hated my guts. And they discovered they hated my guts because they vehemently disagreed with my handling of the the Noordhuis pronoun incident.[1] And on this, I have no regrets -- and will have no regrets.
Hopefully that clears up where the (seemingly limitless) venom is coming from -- with my apologies for dragging confused bystanders into decades-old internet beef!
I (obviously) don't agree with your characterization, but given that I don't work for you (and won't) and that you don't work for me (and won't) can we just let a decades-old disagreement live in the past? Not that it will stop or dissuade you, but I would point out that following me around on HN just to leave nasty replies is exactly the kind of harassing behavior that you so frequently decry in others...
You cannot get upset about someone saying that they would sack you if you were working for them when you did the same to someone else.
I would not work for you because you exhibit all the qualities of a bully. You've been exhibiting this behaviour for decades now. Every time I see you say that you needed to "teach the hardware guys a lesson", or you write a blog post that you would sack a non-employee, or you resort to invective when you could calmly address the issue without going full nuclear, then it confirms my opinion of you.
I'm not following you around HN, incidentally. You just happen to comment on posts that I'm also interested in. I know, however, that if I ever did work for you or worked in the same organisation as you, I'd likely become your target.
I am biased as an ethnic Chinese, but I feel modern medicine is afraid that it's approach, the sum of parts empiricism may be incomplete, in that we dont understand all the parts yet.
The human body is not just human DNA organs working together, but also an ecosystem with myriad bacteria, and we are still in infancy when it comes to understanding the bacteria.
TCM seeks a black box metaphorical approach, which sounds like quackery but I do think it is capable of addressing _some_ blindspots in modern medicine, eg why some medication would work on a yin body but not a yang body... the difference is in the bacterial ecosystem.
That said, I see TCM (and other traditional approaches) as a last resort when modern medicine fails, and I certainly agree the approach is incapable of resisting shamanic beliefs.
Why not apply scientific method to Traditional Chinese Medicine and use double blinded placebo controlled trials to test its validity? In the end also modern medicine is using exactly this method and can treat substances using the black box method.
The fundamental problem is TCM acknowledges individual differences that cannot be measured or even dont exist in the eyes of modern medicine, eg identical twins with different diets will have different responses to the same treatment, so going double blind will mean the results will be inconsistent.
This is why sample size is important tho. With a large enough sample size, you can ignore differences in individuals bc the trend of the control will be smaller than the experiment (or not)
You have no problem accepting eg a treatment can only work on a man, but not on a woman. But modern medicine have no concept of a yin body type and a yang body type, which may or may not be male and female.
The whole idea of TCM is balance, and it varies with the individual, unlike modern medicine, where there is a right and wrong answer to everything. Bacteria bad, antibiotic good. Fever bad, paracetamol good.
Take fecal transplants for example. I dont think it is well understood how it works or it will be a pill by now, and is a last resort when all else fails. And it doesnt involve killing all the bacteria, but restoring balance to the bacterial ecosystem.
Is this yin, yang somehow measurable? If not then there is a fundamental problem.
Also Western medicine is very well aware of side effects, it's actually one of the fundamental concepts. For example it knows that taking Paracetamol is good against pain, but increases risk to the liver, especially when taken with alcohol. It's also very well aware of causes of fever and doesn't recommend lowering it for the sake of it, only from certain dangerous level. It also knows that taking antibiotics affects gut bacteria, so it's often recommended to take also probiotics. It knows that some medicine could affect women differently, especially when they are pregnant or are breast feeding. The list goes on, it's never black and white.
Most religions have the concept of ritual cleanliness for thousands of years, esp touching dead bodies make them unclean and yet at some point, doctors have to be reminded to wash their hands after performing autopsy.
How did we get there? Because "modern science" rejects superstitious beliefs and ritual cleanliness is superstition. Right?
I chose antibiotics and paracetamol as examples precisely because it is well understood _now_ . You go back 50 years before we understand gut bacteria or the difference between male and female bodies and suggest the same, the then modern medicine will laugh at you and call you a witch doctor.
What you claim is simply wrong. The problem was not that the doctors ignored religious superstition, but the Christian customs had been altered over the time and touching the dead was actually act of compassion - there were no impurity laws in late Christianity like it was in Judaism.
I am also Chinese and this is exactly how I feel. The experience of going to a doctor with minor ailments, only to be sent away with the attitude of "take some paracetamol and come back when symptoms worsen" is maddening. In the mean time TCM practitioners have answers that often work for these kinds of things.
In people's zeal to point out TCM's problems (due to its pre-modern scientific roots), I feel like they're also throwing away its potential. Skepticism shouldn't be about wholesale dismissal (which is just intellectual laziness masked as rightenousness) but about improving outcomes.
> In the mean time TCM practitioners have answers that often work for these kinds of things.
You can also take some homeopathic remedies and do a couple of chiropractic adjustments meanwhile. I've also heard that some Christian Science practitioners work wonders if you give them all your Earthly belongings.
The ability to say: "It's likely a viral disease. Wait and see if it worsens" - is a pretty powerful point _in_ _favor_ of modern medicine.
And just who here is rejecting the viral model? Saying "other practitions have stuff worth exploring" is not at all the same as "the regular western medicine model should be abandoned". Why do you feel threathened? This makes no sense to me.
Yet the inverse is not true: the prevailing attitude on HN here is not "western medicine is here to stay as staple but other practitions can add value on top", but to dismiss other practitions wholesale based on their inability to conform to intellectual standards, regardless of measurable outcomes. This is "my god is the only god" all over again.
A chiropractor was able to heal my back problems where months of going to a physiotherapist failed to do so. Aren't we supposed to stay humble and curious for new avenues of scientific exploration, rather than dismissing everything we don't understand?
The unspoken part is the human mind is a big part in health, and treatments that does nothing medically but fools the human mind can work wonders too.
There is a lot we do not understand yet, just as blood letting was conventional medicine a few hundred years ago, and it isnt even entirely wrong since we still use leeches and some treatment, I think we have much to gain if we are not hasty in dismissing alternative approaches.
That said, I fully agree homeopathy and chiropractherapy are full of bullshit and potentially dangerous. TCM, as practiced in a certified scholarly environment in Asia, expects the practitioner to have a considerable basic knowledge in modern medicine too, and is humble enough to acknowledge TCM cannot solve everything. A good TCM practitioner will refer you to a GP when they know modern medicine is more effective.
There are symptoms and idiopathic conditions I've had that multiple specialists at top-tier hospitals were unable to diagnose, but mostly because they were too-narrowminded in their approach (blood tests, etc.) to see the big picture.
I always believed that the AI/LLM/ML hysteria is misapplied to software engineering... it just happens to be a field adjacent to it, but not one that can very well apply it.
Medicine and Law, OTOH, suffers heavily from a fractal volume of data and a dearth of experts who can deal with the tedium of applying an expert eye to this much data. Imagine we start capturing ultrasound and chest xrays en masse, or giving legal advice for those who needs help. LLMs/ML are more likely to get this right, than writing computer code.
Somehow, LLMs always seem to be "more likely to get this right" for fields other than one's own (I suppose, this being HN). The term "Andy Grove Fallacy" coined by Derek Lowe (whose articles are frequently posted here, the term being referenced in a recent piece[1]) comes to mind...
I figured the fallacy you were talking about was the one Michael Crichton describes about reading a newspaper article on a topic he knows about vs one he doesn't, but it turns out that's called the "Gell-Mann Amnesia effect." [1]
> You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward—reversing cause and effect. [...]
> In any case, you read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know.
(My spouse was an ultrasound tech for many years.)
The problem with an example like ultrasound is that it's not a passive modality - you don't just take a sweep and then analyze it later. The tech is taking shots and adjusting as they go along to see things better in real time. There's all sorts of potential stuff in way often bowel and bones and you have to work around all that to see what you need to.
A lot of the job is actively analyzing what you're seeing while you're scanning and then going for better shots of the things you see, and having the experience and expertise to get the shots is the same skills required to analyze the images to know what shots to get. It's not just a matter of waving a wand around and then having the rad look at it later.
This is one of the many places where computer people simplify other professions.
Legally yes, the rad is the one interpreting it, but it's a very active process by the technologist. The ultrasound tech is actively interpreting the scans as they do them, and then using the wand to chase down what they notice to get better shots of things. If they don't see something the rad won't either, so you need that expertise there to identify things that don't look right, it's very real time and you can't do it post hoc.
Did anyone suggest robots do ultrasounds? Who is simplifying it? Having literally just done one: the tech came in, basically said nothing and took a bunch of pictures and then Doc came in and interpreted the results.
People are suggesting that AI interprets the images, which is a fundamental misunderstanding of the process, because the tech is making choices while taking the pictures of what the images should be of. You can't wait until the pictures are taken and given to the rad before the interpretation can begin, it has to be happening during the whole process. The question then is what is the place of the AI in that process? What is it automating?
At my spouse's primary clinic the techs did debrief the rads in person, and at other sites they include notes with the images, so they do get supplementary information from the techs. I suppose AI would make the rads superfluous because between the tech and the AI they don't have anything to do, but the majority of the effort will remain, and I still wouldn't trust it giving AI's predilection for just making things up when there's missing information.
When AI writes nonsensical code, it's a problem, but not a huge one. But when ChatGPT hallucinates while giving you legal/medical advice, there are tangible, severe consequences.
Unless there's going to be a huge reduction in hallucinations, I absolutely don't see LLMs replacing doctors or lawyers.
100% agree ‘chat bots’ will not be a revolutionary technology, but other uses of the underlying technology will be. General robotics, pharmaceuticals, new matter… and eventually 1st line medicines and law sure, but I sure don’t want doctors to vibe diagnose me, or lawmakers to vibe legislate.
LLMs have a history of fabricating laws and precedents when acting as a lawyer. Any advice from the LLM would likely be worse than just assuming something sensible, as that is more likely to reflect what the law is than what the LLM hallucinates it to be. Medicine is in many ways similar.
As for your suggestion to be capture and analyze ultrasounds and X-rays en-mass, that would be malpractice even if it were performed by an actual Doctor instead of an AI. We don't know the base rate of many benign conditions, except that they are always higher than we expect. The additional images are highly likely to show conditions that could be either benign or dangerous, and additional procedures (such as biopsies) would be needed to determine which it is. This would create additional anxiety in patients from the possible diagnosis and further pain and possible complications from the additional procedures.
While you could argue for taking these images and not acting on them, you would either tell the patients the results and leave them worried about what the discovered masses are (so they likely will have the procedures anyway) or you won't tell them (which has ethical implications). Good luck getting that past the institutional review board.
AI code assistants are amazing when you start from zero and just need a 80% working prototype. But once you start trying to refine the product from there, that's where the automation gets counterproductive.
If you can exactly specify the problem, eg "Password input crashes when the password has an apostrophe", AI can probably fix it. But if the bug report comes in as "Password input randomly crashes", I will be very surprised if AI can figure out why and fix it.
Where a human wrote the code, he or she may figure out why fairly quickly. Now, if you want a human who didnt write the code to understand the AI generated code, it may take a lot longer.
In fact, in all likelihood, the AI assisted products are likely to be buggier and stay so longer, esp if companies start to think they can fire the senior devs and hire less skilled devs and fill the gap with AI.
At some point, the pendulum will swing back, and companies will be chasing devs again.
It's certainly going to be fun seeing this all play out over the next 10 years.
In some ways it's like taking over a project written by someone else that's "80% done". You're locked into their design, get to analyze all that code 1 bug report at a time, get frustrated by obvious mistakes, confused (and mislead) when they relied on some clever side effect.
The quality of life in this maintainence mode depends enormously on the quality of the original coder. "Why did you choose this over that?" Is a common question I have for earlier devs. The AI answer is the least satisfying "it seemed like a good probability at the time".
IME writing the app from scratch to "done" is 10% of the lifecycle of the code. It's the other 90%, spanning over decades, where the quality (or lack thereof) reveals itself.
Personally I'm finding AI useful as a tool. Would I want to be the human fixing AI bugs? (From human bug reports which are pretty vague?) I'm not so sure about that.
As someone who was working on Itanium systems at the time, I can confidently say that "consumers" never had a chance to choose Itanium. It was only available on server hardware and Windows ... will there as a version for it but it was slow ... all the compatibility and performance downsides of M1 Macs, and more power hungry too... practically no upside.
It is still a common belief Intel created the Itanium to fool the proprietary unix vendors into thinking they can stop developing their own RISC CPU, but as soon as it happened, Intel gave up on Itanium as well. Solaris lasted a little bit longer only because Sun gave up early on Itanium.
Google suffers from classic Innovator's Dilemma and need competition to refocus on what ought to be basic survival instincts. What is worse is the search users are not the customers. The customers of Google Search are the advertisers and they will always prioritise the needs of the customers and squander their moats as soon as the threat is gone.