Yeah, one of those machines (either the Palo or the Alto) was put on display near the entrance of the late Xerox PARC on Coyote hill (near VA). I was playing with the idea to power it but was told that it doesn't work anymore, at least the power supply is dead and allegedly there were also missing components... I suppose SRI will toss it soon if not already...
Never heard of the Xerox Palo computer and I worked at Xerox for much of the 80s. Xerox made a lot of different machines (mostly but not exclusive in the D-Machine family) but if there was a Palo machine I’d be interested in seeing a photo.
I love this mini article, however, I disagree with the main conclusion that collaborative editing is not an algorithmic but a UI/UX problem. I think that collaborative editing is a semantic problem. To the best of my knowledge (I'm writing this comment without much preparation), all SVN/Git algorithms are based on UNIX diff (Hunt–Szymanski algorithm). UNIX diff (and patch) is purely syntax driven.
Actually, I will make a small deviation here: I think it is a big industry/startup/open source project there, in creating a set of semantic diff algorithms/implementations. For example, due to my present job, I am very interested in collaborative editing of electrical circuits, and layouts for PCB and chips. Altium and KiCad are trying, for exmaple, to store everything in XML/text files and put the text files in Git/SVN and I can tell you a botched C++ program is nothing in comparison to a botched and malformed electrical circuit. So we need diff tools that "know" about a text file, vs rich-text with formatting, vs bitmap vs vector image, vs song, vs English text. Anybody want to start an open source project (DM me or put a comment here).
Anyhow, thanks to the authors on the great insights and let's work on the take home!
What an obvious article. But it, because it comes from Apple, everybody pays attention. Proof by pedigree. OK, here is my two cents. Firstly, I did my Ph.D. in AI (algorithm design with application to AI) and I also spent seven years applying some of the ideas at Xerox PARC (yes, the same (in)famous research lab). So, I went to and published at many AI conferences (AAAI, ECAI, etc.). Of course, when I was younger and less cynical, I would enter into lengthy philosophical discussions with dignitaries of AI on what does AI mean and it would be long dinners and drinks, and wheelbarrows of ego. Long story, short, there is no such thing as AI. It is a collection of disciplines: the recently famous Machine Learning (transformers trained on large corpora of text), constraint-based reasoning, Boolean satisfiability, theorem proving, probabilistic reasoning, etc., etc. Of course, LLMs are a great achievement and they have good application to Natural Language Processing (also intermingled discipline and considered constituent of AI).
Look at the algorithmic tools used in ML and automated theorem proving for example: ML uses gradient descent (and related numerical methods) for local optimization, while constraint satisfaction/optimization/Boolean satisfiability, SAT modulo-theories, Quantified Boolean Optimization, etc., rely on combinatorial optimization. Mathematically, combinatorial optimization is far more problematic compared to numerical methods and much more difficult, largely because modern computers and NVidia gaming cards are really fast in crunching floating point numbers and also largely that most problems in combinatorial optimization NP-hard or harder.
Now thing of what LLM and local optimization is doing: it is essentially searching/combining sequences of words from Wikipedia and books. But search is not necessarily a difficult problem, it is actually an O(1) problem. While multiplying numbers is an O(n^2.8 (or whatever constant they came up with)) problem while factorization is (God knows what class of complexity) when you take quantum computing into the game).
Great, these are my 2 cents for the day, good luck to the OpenAI investors (I am also investing there a bit as a Bay Area citizen). You guys will certainly make help desk support cheaper...
I don't think I'm alone in reading "they ** up" as "they fucked up" in my head, without meaning to. So whoever would be offended, is still offended, only now they can't complain about it since the person technically didn't swear.
There's no stakes here (the internet) for it to matter if you swear or not.
Your attitude on the internet won't result in any meaningful feedback, by default.
Someone has to be watching from a more strict environment for it to matter what you say on the internet. It needs to affect your job, or your wife needs to read your posts, or a moderator needs to clamp down on you.
The community accepts swearing by default.
Soo people who are in professional environments are going to bring their communication norms with them when they post online.
An I over
-explaining an answer to a rhetorical 'why?' question. I think I am.
I can never tell online, if people are literally asking a question that requires explaining OR it is simply a rhetorical question.
This is a great diverging discussion about the "censorship" and the commenting style in public forums. Let me add something: when I was writing the original comment that started all this, I wanted it to be funny, catchy, snarky, and karma-point-worthy. So, I decided to put a swear word as a stylistic choice. Then, to add some extra to the short comment, I decided to poke fun at the whole FCC bleeping regulation. Here, I will warmly recommend a book dedicated to bleeping over the seven "unspeakable words": the famous Steven Pinker's famous book "The Stuff of Thought".
Now, about the number of asterisks: I really didn't think that much about this, I kind of think I counted the right number of letters, but then there is no caret overwrite mode in the bleeping browser, which is yet another story...
Well it's pretty clear I injected my own world view onto your comment and enjoyed it.
The FCC bleeping regulation doesn't really interest me, it's mild irritant in media..
It's slowly becoming obvious to me that the English-speaking cultures (UK, US, AUS, CAN, ect) are sculpted to interpret the same content in their own cultural view and it all somehow works even though we're constantly talking past/parrallel to each other.
Watching an American historian's reaction to philomena cunk peice, showed me how much he enjoys the literal-ness of her straight faced jokes, the same jokes I see as boldly satirical comments made for chuckles. And it somehow works for both cultures.
With that in mind, communicating anything worthwhile over internet comments is obviously orders of magnitude harder because of the cultural barriers. And the cracks between worldviews run deep.. and are starting to show.. I don't really know how to comment to other cultures accurately anymore.
Google helping with your security is similar to when those nice mafia guys knock on your door offering protection. Don't forget that Google is apotheosis of evil corporation trying to take over all your data. This is the very company that turned "don't do evil" into "do things".
It's almost like a giant company like Google have ~100,000 employees, with a complex incentive structure at different levels that are encouraged to do different goals. While I have almost completely de-googled my life a few years ago, it is just stupid to attribute malice to anything they touch -- they have plenty of good contributions, certain parts of AOSP being an example.
I would be happy to have that on a GrapheneOS phone for example, if I hadn't went with Apple.
This is not a Googler's 20% project. You can expect any major feature released to have been infected with Google's morality and mentality of being an ad company needing to extract profit out of free products.
You cannot apply Hanlon's Razor to megacorporations.
I think a lot of Android devices have that; you type ##4636## into the dialer, and a menu appears. In that menu, you can select which cellular technologies are used.
I wish there was a way to create or apply a rule that addressed that type of comment. Most, certainly not all, end up not contributing much, just angst.
Yes? The world is nowhere near perfect, but those mafia guys are probably actually going to protect "their" money from other gangs, and Google's obsession with your data means they have even more incentive to protect said data from other actors. In addition, of course, to the more general incentive to build features that can make more people (or in this case, organizations) choose Android.
Credit where credit's due. Google moves against our best interests very often but this is not one of those times. Let's accept this improvement graciously. Other Android-based operating systems like LineageOS and GrapheneOS will also benefit.
One shall not obsess with the game controller they use, I suppose, in their analysis, failing the controller wouldn't prevent the sub from resurfacing... It is symptomatic, however. The whole Silicon Valley is like that. You have the old guard, NASA, millions for fault analysis, rumors go the C code on some NASA rocket was certified at $1000 per line---human eyeballs reading it...
All those start-ups and unicorns, though. Just do it, when to be a hacker became an honorific, a title... Code your stuff, design your mechanics at Starbucks, who needs mathematics and physics when we can have s*t done instead. Who needs signing-off when we can have a carate-belt meeting standing on bouncing balls instead...
I don't want to be the dean of the faculty of prophecy but this is only the beginning. Wait until those autopilots starts using ChatGPT/Stack Overflow/copy/pasted code...
The use of the controller is symptomatic of their attitude, but I agree that it's probably not as relevant as, say, using an untested (and apparently untestable) carbon fiber hull:
> Lochridge’s concerns mainly focused on the company’s decision to rely on sensitive acoustic monitoring – cracking or popping sounds made by the hull under pressure – to detect flaws, rather than a scan of the hull.
Lochridge said the company told him no equipment existed that could perform such a test on the 5in-thick (12.7cm-thick) carbon-fiber hull.
“This was problematic because this type of acoustic analysis would only show when a component is about to fail – often milliseconds before an implosion – and would not detect any existing flaws prior to putting pressure on to the hull,” Lochridge’s counterclaim said.
Interesting... This is similar to the concerns that Vince Weldon, the Boeing engineer had over the 787 Dreamliner hull made of composites: that flaws could not be found by a visual scan.
I have a carbon fibre fork on my bike. I had a low-speed accident a few years ago, and took the bike to a mechanic to get it checked out afterwards. He said he'd checked everything, and it was fine - except the fork, which he had no way of telling if was undamaged or about to fail. I think i'll get steel next time.
this isn't true anymore. you can get carbon bikes scanned after and accident to find any internal damage. Ruckus in Portland, OR is where i've shipped to before with luck, but there's a lot more local options now depending on the scene
Ouch! Folks across the street own a bike shop and sponsor an annual criteruim. Last year there was a crash right in front of our house - touch of wheels I suppose. There's a slight hill at the end of the street and I guess the riders could have been going in excess of 30 mph or more. No one badly hurt AFAIK but one bike had a broken frame. In this case no fancy equipment was required to detect the flaw. :-/
You may be interested in having a look at this relevant and interesting article "Nanoscopic origin of cracks in carbon fibre-reinforced plastic composites":
Xbox controllers are way ahead of low-end logitech controllers though, so the point stands: If you want to control the device using a controller, might as well pick a good reliable one instead of the cheapest option available.
I worked for a defense contractor, and we had initial plans for a contractor to build us a wireless controller. Their controller was an ugly box-shaped controller that they wanted to charge us 200k. I redesigned the architecture of our system so that we could use an Xbox controller instead.
There is no reason why these controllers can't be used over some other "industrial" controller, as long as you design the system appropriately to account for failures, which you should be doing regardless of what controller you select.
I’ve always hated this aspect of journalism- news stories latching on to an innocuous detail and then playing it up to make the whole operation seem like a circus. It plays to the Dunning-Krueger effect so hard.
Nothing is untestable. If validation results in destruction, then destructive testing is how this is done.
Failing to test is a choice, similar to the choice made by a software engineer who determines, without testing, that their production service has no bugs.
If a component's flaw can only be detected right before a failure, we need statistical analysis of multiple such failures being intentionally induced to estimate the component's limits.
Not finding or knowing about existing limits or flaws is the direct result of not looking for them.
As a Quality Assurance person, this is what drives me up the wall, across the ceiling, and down the other wall.
You can always do a test run. Not doing that test run is essentially saying every person you subject to that system is not worth the trouble of doing the test.
No one was claiming that it was untestable, the company did decide to go with an acoustic monitoring, which is a destructive test, instead of a scan, which is a nondestructive test. The company claimed there was no non-destructive test that could be done. The issue is that when you're dealing with short run custom components, like the hull of a submarine, just because one passes a destructive test doesn't mean the others are free of manufacturing defects. You can't do useful statistical analysis when your sample size is small.
The details about the single window on the craft is also pretty damning:
>Further, the craft was designed to reach depths of 4,000 meters (13,123ft), where the Titanic rested. But, according to Lochridge, the passenger viewport was only certified for depths of up to 1,300 metres (4,265ft), and OceanGate would not pay for the manufacturer to build a viewport certified for 4,000 metres.
I read about the huge lack of safety measures on the submarine, and now this. I am beginning to think that perhaps the CEO of the company has voluntarily decided to die with his unfortunate fellow passengers. I know mine is big speculation, but taking the glass viewport to 300% of its certified maximum strength in an environment where the slightest setback can be fatal can no longer even be considered as "taking risks."
He says that testing of the composite materials is really difficult in comparison with metal. As they fail catastrophically without material stress.
Quotes like: "Lochridge’s concerns mainly focused on the company’s decision to rely on sensitive acoustic monitoring – cracking or popping sounds made by the hull under pressure – to detect flaws, rather than a scan of the hull."
For the most part (so no self-driving cars), startups are in a low-stakes game, so it makes sense to hack away, and it nas nothing to do with an "old guard." Google (or any other tech company) outages aren't particularly memorable unless you were an engineer involved in it. The space shuttle accidents were labeled "disasters."
> Theranos is an outright fraud. Not the product of getting things done on fast track
Theranos is product of the same culture. Company founder chases a goal they do not understand, and convinces investors its right around the corner. In actuality they have no clue.
There is no foundamental difference between theranos diagnosis and tesla autopilot.
The only reason we call theranos fraud is that we know conclusively that their goal is impossible. We have not yet accepted that autopilot is.
Elizabeth Holmes was not convicted of fraud and sentenced to eleven and a quarter years in prison because her company was a bad investment, no matter how much tech "journalists" want you to believe. She's going to prison because she committed boring financial fraud. If she was truthful to investors about the state of her company and her technology, she never would have received funding. Claiming that your technology will totally be ready in 6 months and will change the world is one thing, saying that it's in secret use by the US Military in Afghanistan onboard medevac helicopters is another.
> we know conclusively that their goal is impossible
No we only know that their approach was wrong. Here is a preprint of a paper describing detection of 12 different types of cancer from a single drop of blood:
The problem isn’t the volume of blood, it’s that blood drawn from finger capillaries is not a representative sample for diagnostic purposes. There are plenty of tests that are even more sensitive, but they require a proper sample drawn directly from a vein or artery.
Really? Think another way: MVP. MVP is a big in the startup scene. No startup tells you how they made there MVP. Most of them think, if I get enough investment, if I get enough customers, then I will make it better. But until now the other way is just enough for my MVP.
So some would call it MVP and others would call it fraud. Until it is really being implemented as advertised.
I agree. With software being ever more embedded in our daily lives and ability to function in society, the stakes are very high, even in domains outside of physical products.
Social media has shown to be pretty destructive. I don’t think there are many low stakes areas for software. It’s all impactful and potentially harmful at scale.
As far as I remember the Theranos people used standard commercial tests, and sold the results as coming from their magic device. So the results were solid - the fraud was elsewhere. I might be remembering this wrong.
No expert, but as I have read and understand, theres a catch. Theranos promised all kinds of tests drawing minimal blood, and they performed the tests using commercial equipment but with less blood than what these equipment is meant to use. So the results were not actually solid.
Like the video [1] mentions, they were only improvising on things (including the controller) that would not pose a safety hazard if they failed. The critical components like the capsule were designed alongside NASA and others. People aren't just tossing safety to the wind, but trying to create a better balance. NASA does spend a extreme amount of money on compliance and safety - yet that doesn't prevent them from doing things like blowing up $600 million Mars probes because of a mismatch between Imperial and Metric units. [2]
Basically you cannot, no matter how much money you spend, prevent every possible mistake, or even every "obvious" mistake, because "obvious" is often only obvious in hindsight. And going too far on the side of risk avoidance leaves you frozen in time, unable to progress, even as you continue to make mistakes - which drives you even further into extremes of risk avoidance. Of course on the other end being completely cavalier about safety leaves you making mistakes you both can and should have foreseen.
So I suppose we'll just get to see which this was. If anything my prediction here would be that they started becoming so comfortable with these dives that they impacted the Titanic, going for that epic view, resulting in a cascade of system failures or even a breach, bearing in mind you're already going to be near critical pressure thresholds. Absolutely zero basis for my prediction, but I think it's much more probable than a controller failure. They had redundancies on the controllers, and could surface without them. But human hubris has no such constraints.
He told CBS News: "You know, there's a limit. At some point safety just is pure waste. I mean if you just want to be safe, don't get out of bed.
Don't get in your car. Don't do anything. At some point, you're going to take some risk, and it really is a risk/reward question. I think I can do this just as safely by breaking the rules."
I’ve know mountain climbers, hang glider/paraglider pilots, scuba divers, sky divers, and more and we pretty much say that exact thing. I’ve been casual acquaintances with people who have died doing extreme sports.
Usually, the people in these sports say if the worst were to happen to them, they accept that risk and wouldn’t want their fate to stop other people from following their passions. That the feeling of being alive chasing these dreams fulfills a part of them, and they would be lost without that fulfillment.
We all do take risks for our own happiness, different people just want different things from life and will drawn that line in different places.
My guess is the five people on the sub would not want this incident to stop people from attempting to visit the Titanic.
...Yes, you're going to take risks. Got it. This risk in particular ends up with you as chunky salsa 4 kilometers under the water, because you couldn't be bothered to sink a copy of the bloody thing with no one in it first.
If it's worth making, make 2 or 3 of with at least one slated for being destroyed at one point.
Space Shuttle Enterprise never saw Space or the launch pad. It was still made and flown.
There's a spectrum of risk comfort in mountaineering though.
At the end of the day if there's a rockslide, there's nothing you can do, so you have to be willing to accept some level of risk.
But at the same time, the majority of mountaineering deaths are due to people taking unnecessary risks (for the sport). Being poorly prepared in terms of gear, or making poor decisions, or not being educated enough in climbing, safety assessment, and rescue techniques.
There's absolutely a lot you can do to mitigate the risk, and I think even the experienced mountaineers who choose to take more risk (say, traveling lighter and placing less protection in order to cover more ground in a day) don't advocate for that as the standard way to practice the sport.
I'll be honest, I've had a few close calls myself, but I've taken those as lessons in how to do things more safely going forward, because I value doing things safely more than "bagging more peaks".
I guess if your dream is something like "climb every mountain" accepting more risk is necessary.
But I don't see why the people on this Titanic expedition couldn't have achieved their goals while spending a bit more time on R&D + QA.
Also, people should be making these risk decisions for themselves, not others. The CEO may have had it coming to him, but I feel bad for the other crew on board if they weren't able to decline the expedition, or weren't provided adequate information about the testing that had been done.
This is not true, the maker of this sub compromised substantially on safety and did not have it certified by an independent body like most other builders of deep sea submarines do. In fact, members of the small deep sea submarine community wrote a letter to the company warning them about the safety hazards and the lack of certification. Source: Today's interview with a deep sea exploration submarine commander on German TV (ARD Tagesschau).
In particular, this company used carbon fiber materials that were neither tested nor certified for the intended operational depth, and they also did not do extensive testing like you do when you certify a craft.
As it's explained in another comment - the CEO wanted it tested but there isn't a facility anywhere on earth equipped to test it. Perhaps the argument then is that part of the company funding should have been building a custom rig designed for testing this highly custom 5 inch thick carbon fibre shell - but it's not like options to test it were available and were ignored.
This is not true either. As the submarine commander interviewed lays out, the testing is done by actual dives, and there is a certification body for it. Almost every other deep sea exploration submarine is certified. However, getting certified is voluntary.
Of course, certifying it just means that independent experts take a look at the design and use sensors during dives. All of these vehicles are experimental. But not even trying is negligent.
> the CEO wanted it tested but there isn't a facility anywhere on earth equipped to test it.
They could have dropped it to the bottom of the sea with weights, have a timer release the weights and hauled it back up. Cheaper than renting any testing facility on earth.
They could have done this multiple times on the same hull until it imploded. If a month of dropping it and retrieving it doesn't result in implosion, then it's probably safe enough to put people into it.
There was no need to make its first test of that pressure with people in it.
A lot of things that we use all the time couldn't be reasonably tested in operational scenario in any other way than by just using them(most planes are in that category - you can test a lot of things on test benches but there isn't a way to test an airliner at operational speed an altitude other than actually flying it)
I think if SpaceX can send 15 robotic rockets into space to test them, this lot could've sunk a prototype to high pressure and got a diver to go down and smack the hull with a wrench for a while.
>Like the video [1] mentions, they were only improvising on things (including the controller) that would not pose a safety hazard if they failed.
This is an extremely dangerous attitude though (obviously). In the BBC documentary, the thrusters are out of orientation and they just solve it by saying 'turn the controller, since right is your new forward.'
In the documentary it's played off as fun, but there's a non-zero chance the sub is currently stuck within seafloor wreckage and it could 100% be the fault of the controller.
Bad steering certainly wouldn't prevent them from dropping ballast (which they imply is the ultimate failsafe), but it is not a zero-failure-risk system.
The critical components like the capsule were designed alongside NASA
I saw that claim, and wondered what it entailed. The founder is clearly in the mold of testing mission-critical/life or death systems in production, while skirting or floating the rules. Such a claim is like one of those logos you see on the front of software start ups home pages, claiming some usage by Google, MIT, or CERN, without any verification or context.
Their release is here [1]. The NASA tracking number is SAA8-2031655. The agreement started in 2020, and concluded in 2022. One of the guys on board the sub is Stockton Rush, the founder/owner of the company. He clearly believes (or believed as it may ultimately be) in the product they created, and wasn't just casually playing games for bucks.
I would largely tune the media out on this. They have no more special insight than you or I. All that's known is that a sub has gone dark; everything else is speculation. Within 3 days we'll know for certain whether the people on the ship have been rescued, or are dead. In case it's the latter, we won't have any realistic idea of what happened until the sub, or its remains, are recovered. And that may simply not happen.
Why is the capsule considered the only critical safety component?
If everything else fails wouldn't it just fall to the bottom of the ocean? Even if it can survive that depth that only sounds good on paper. Nobody could get to you.
Because an implosion is a single point of failure leading to instant death. Anything else can be salvaged by staying calm and using one of multiple ways to resurface.
Subs will always have 1000 ways to kill you and most of them are catastrophic. You're pretty much boned if anything goes wrong especially with so few crew members, lack of compartments, backup propulsion, etc.
The reason NASA did that is because the stakes were so high. They couldn’t send a firmware update in the middle of a journey and the reputation of America hinged on the success of NASA.
If you are under that level of pressure, you make sure you dot your i’s and cross your t’s.
I'm really frustrated by anyone that poo-poos this specific issue. One shall absolutely fucking obsess, me for instance about the game controller they used. Look at this.
So the F710 Logitech gamepad was something I wouldn't even trust as a teenager to let me play emulated retro games accurately. It has a hardware switch to toggle between DirectInput (an ancient input technology within Microsoft stack), and XInput which is a non-trivial third-party software to map controllers to do things on your computer. Complaints range from 'the input switch was finicky and I had to tape it down' to 'you had to insert the USB dongle into the computer carefully like trying to get a Nintendo cartridge to work'. In this case, it was the main navigation control on a submersible far exceeding its safety buffer by orders of magnitude.
Sure, military applications have found use for XBox and other controllers as system inputs because generations of those serving are used to that interface. The differences are not subtle:
1. There had to be a certification process to use the controller, including risk analysis which would absolutely deny any attempt to use wireless controllers to operate any piece of equipment, no less anything that was weapon or safety critical. The military doesn't drive boats with XBox controllers.
2. XBox controllers are well designed and reliable, at an order of magnitude of this one. THEY ALSO COME IN A WIRED VERSION.
3. They decided to further mod the controller for no reason with longer sticks, maybe to make it look less like the toy it is which actually reduces the dead spot response of an analog joystick. Why? What the fuck.
The reason for the photo above is to show what I've seen in every media piece where they talk about it: It is NEVER wired in any video or photo. In one video, which I can't dig up right now, the CEO actually tosses the controller carelessly on the ground and quips that "these things are designed for 16 year olds to throw them around".
Edit: The bit mentioned above is at 0:37 in this video: https://youtu.be/ClkytJa0ghc?t=37 - "we keep a couple of spares on board just in case", that's great but I'm wondering who in that sub knows how to pair an ancient Logitech gamepad and remap XInput like a kid playing Diablo.
That is on video and a joke during the interview, forget for a second that they're using a controller from 2005-era for a moment; at that point on something like an airplane or a battleship, you just caused an incident that requires maintenance testing to make sure you didn't just disable the fucking vehicle by throwing the controls around for fun.
No. Don't disregard this specific reckless decision even though it likely had nothing to do with whatever went wrong in this situation. The CEO of the company from the get-go abandoned any semblance of basic safety in engineering, and fired/sued the guy in charge of safety when he called it out. This controller is a metaphor for something that needs to immediately be regulated so that people like this can never get away with it again.
. Edit: The bit mentioned above is at 0:37 in this video: https://youtu.be/ClkytJa0ghc?t=37 - "we keep a couple of spares on board just in case", that's great but I'm wondering who in that sub knows how to pair an ancient Logitech gamepad and remap XInput like a kid playing Diablo.
They don't, in the BBC Travel episode "Take Me to Titanic" about this team and their sub, they did a dive with customers and almost had to abort because the controllers stopped working.
Then, when they reached the bottom of the sea, they were stuck going around in circles on the sea floor because the keymappings on the controllers stopped working. They almost had to abort the dive until someone on the surface had troubleshooted the issue and found out that the keys were incorrectly mapped to the sub's controls.
It turned out that the 'left' direction key on the gamepad actually mapped to the 'forward' motion of the submarine, and all of the gamepad controls were incorrectly mapped.
They only found out about the incorrect mapping after a dozen hours underwater and once they reached the seafloor.
That's just the insanity that happened with the gamepad, the entire dive experience was littered with potential and actual emergencies that were only averted due to luck.
To find their way around, they were literally dead reckoning on the bottom of the seafloor without any navigation systems. They had no voice or video communications with their team on the surface, and they relied on that team's text messages to help them adjust their dead reckoned position.
Watched this too. I am alarmed by the CEO's assertion in the CBS video segment that only the hull was critical, everything else could break and you'd be safe. We know for a fact that other submersibles have gotten caught in wreckage down at the site. Being able to control your thrusters and steer could be the difference between life and death.
I'll admit I absolutely don't have this adventurous spirit, and I admire those who do. I'm sincerely praying for their successful rescue. But some of this does seem careless and possibly negligent.
> This controller is a metaphor for something that needs to immediately be regulated so that people like this can never get away with it again.
There is something to be said for protecting people who would sign up for such an experience while misjudging the risk, or assuming that some government agency does provide oversight. Though, customers seem to have been informed that this is not the case: https://twitter.com/FridaGhitis/status/1671120043126423553
But apart from that, if someone wants to build a lego submarine and use it to dive to the Titanic, they should have every right to do so. It is their risk to take. Asking others to pay to join does not change this moral calculus. And yes, the people left on the surface have an ethical responsibility to not unreasonably refuse to rescue them when things to awry.
Sorry but that doesn't work in a reasonable society. You can't let just anyone do anything they want with the repercussions being 'they don't get rescued'. A few points:
* You can't let rescue services decide who deserves to be rescued -- for so many obvious reasons it would be patronizing to list them
* People with a lot of money and no sense can fuck things up on a massive scale if we let them do whatever they wanted
* Human society is not a free-for-all. As much as the civil libertarian tendencies in me want to say 'sure, do whatever you want, just don't mess with anyone else's stuff', it really isn't that simple
Shame and societal norms are a big deal in keeping people in check. Just getting yelled at in public for doing something objectionable is enough to keep most people from spitting indoors or pissing on crowded subway platform or what-have-you.
When shame doesn't work we rely on laws. Laws must be universally enforced and they must be fairly enforced and they must be seen to be enforced.
If we allow idiots with stupid ideas to get lucky enough times then they become looked up to and the shame goes away.
To the same end if the shame doesn't stop them we need to physically stop them or take away their ability to do the societally harmful thing they want to do.
People need to re-learn that you should be embarrassed for failing when what you strove to do was stupid and destructive.
In the general case that is more or less accurate. There are however a bunch of exceptions in the extreme case.
Rescue services do have a point where they will decide not to continue. During the Tham Luang cave rescue (the thai football boys that got stuck in a cave) there was a period where the rescue services decided that continuing was just too dangerous. It was only because of a handful private cave divers was crazy enough to try a exceptional dangerous idea that those children got out there alive, and had it failed then those cave divers would have basically received all the blame.
The case do illustrate how far into the extreme we have to go. The local rescue services gave up and gave the job to nation service. The national service gave up and gave it to the military (with international support). The military gave up, and then through almost a backdoor, a few individuals tried a Hail Mary attempt which against all odds worked well enough to get everyone home.
There are activities where people has to accept that rescue is limited or zero. If you go hiking in no man land then there is a real risk of rescue service not being able to locate you. People who attempt sailing around the world has the risk of being "lost to sea". Cave explorers both dry and wet has to accept that rescue attempts are done based on what is feasible. Same goes for wreck divers.
We could argue that those risky activities should be illegal (or shamed), but the counter argument is that a lot of activities are just inherently risky. Sports generate a huge amount of injuries. Motorcycles are viewed by health professionals as organ donor generators. Extreme sports are extreme, but they tend to also follow more rigorous training and certification in order to address those risk.
> We could argue that those risky activities should be illegal (or shamed), but the counter argument is that a lot of activities are just inherently risky.
There are always going to be distinctions. We allow motorcycles but we don't allow motorized wheelbarrows, or unicycles that can go 60mph. Why? History and tradition, practicality -- whatever the reason, it doesn't necessarily have to make sense and it surely wasn't designed that way. No one makes the standards to which we hold people -- but we are allowed to complain when we don't feel they are in line with a healthy society.
A hand-guided mecha-wheelbarrow would be incredibly helpful for people who would like to do more of their own gardening and landscaping but who don't have the physique for it. I'm now going to spend half a day thinking about how something like that could be built to work reliably, economically, and safely, while other people on HN have probably already spent half a day thinking about how to prohibit it.
Then, back to that unicycle thing. Your "bad ideas" make fascinating engineering challenges. Gotta give credit where it's due.
You obviously weren't paying attention to anything I wrote besides looking for things to criticize. I was saying that those things are all bad ideas but we allow some and not others for reasons that go beyond 'bad' and it is not for us to judge or change.
It doesn't seem that difficult. Get an electronic speed controller and a hub motor from a trashed e-scooter (10" wheels should work?) put a battery pack on it and rig the throttle to a cable on the handle.
I suspect this loss will be as effective as any regulation. Nobody will get on board one of these subs without doing a lot more due diligence than the present customers did. And if they do, well, that's on them.
Believe it or not, new laws aren't always the answer. You can't bubble-wrap the world and you will make it a worse place for everyone if you try.
I didn't say anything about making new laws. I am proposing that our society is heading in the wrong direction by lionizing people who take stupid risks and win.
Yeah, I'm mostly addressing truemotive's knee-jerk post above.
Let's at least wait for the failure analysis reports before calling for congressional hearings. In other words, don't send a lawyer to do an engineer's job.
Life is becoming very safe, but we evolved under conditions of deadly risks around every corner. What we can and should ask is that the risk takers only jeopardize their own lives.
That’s nonsense. “Reasonable society” has been building and using commercial submersibles going on nearly a century and suddenly, after less than half a dozen people die in a dumb vanity stunt, we suddenly need to create a suite of brand new regulations? Our legislators and regulators have much better things to spend their bandwidth on.
What’s next? Are we going to start patrolling for DSVUIs, breathalyzing submarine captains at random?
I'm saying it can be enforced the same way it is in most industries that rely on dangerous equipment: via insurance policies, civil lawsuit, and criminal prosecution for negligence.
We don't breathalyze pilots when they board the plane, despite them being responsible for hundreds of people at a time, and millions of flights complete every year just fine.
No one suggested doing mandatory DUI checks except you, in order that you can argue against regulation. Regulation is requiring the pilot to be sober. Mandatory breathalyzer checking would be an example of a type of enforcement (which no one is arguing for).
They decided to further mod the controller for no reason with longer sticks, maybe to make it look less like the toy it is which actually reduces the dead spot response of an analog joystick. Why? What the fuck.
I would assume to allow more precise control - longer sticks will give you a smaller input for the same amount of travel and require more travel to achieve the same input.
Yeah, I don't get the "eh, it's not that big a deal" reaction.
That it was a game controller? OK, not necessarily a big deal. Hopefully they'd take two of them just in case, even with a high-quality game controller, but still, might be OK.
That it was a game controller not from one of the major console manufacturers? And a cheap one, at that? I can only assume people not taking that as a serious WTF haven't used many of those lately (by which I mean in the last 15+ years). If you buy from any but a handful of trusted brands, these days—and Logitech ain't one of them, it's almost, but not quite, entirely the three console manufacturers, and the exceptions are not cheap—you can expect the device to suck from day 1 and to die completely within a year, even if subjected to no abuse. You may luck out and it won't, but if you deviate from that guidance, you're very likely buying garbage unsuited even to home use.
If someone told me they bought a $30 Logitech controller for their Playstation emulator box, I'd tell them to buy a Playstation or XBox controller instead when it breaks.
This all will be better discussed after the recovery or memorials or whatever (Godspeed to them, I honestly hope for a miracle)
This metaphor is extremely double edged, especially in Silicon Valley ethos. When it fails, everyone piles on about the game controller and the lack of ‘flight worthiness.’ When it works they are lauded as heroes because of the extreme savings of using proven COTS products. The drone pilots are a great example, some officer got a nutty idea to try a game controller as his command had “more experience” with them and their success rates improved; now we are talking about it.
Space and the deep sea are the extreme limits, I think ‘regulation’ is really difficult because it is risky no matter who does it. In a way, this is the ultimate regulation that is going on here. And we still don’t know what happened or what could have failed yet.
When (most) people talk about the shitty controller, the options in their heads aren't "3rd party Logitech" or "$10,000 bespoke controller". They're "$200 elite OEM controller", "$60 standard OEM controller" or "$30 notoriously bad piece of shit Logitech".
>Space and the deep sea are the extreme limits, I think ‘regulation’ is really difficult because it is risky no matter who does it.
But this is why regulations exist in the first place. That's like saying we shouldn't regulate surgery, because it's always risky.
We should require certain credentials and safety measures for people who want to take civilians to the ocean floor in a tiny sub. I don't think that should be a controversial take.
The onus to make sure that a service they are going to use isn't any more dangerous than it has to be, shouldn't be on the consumer. I don't want to live in a society that thinks it's acceptable to hand-wave death and suffering away as if human lives are acceptable collateral in the "free market" correcting itself.
Customers would still understand that the thing they're going to do is super dangerous, but they should be able to rest easy knowing that the company they're using for this meets some minimal level of safety.
Regulation in areas of extreme innovation is by definition difficult. When you have people onboard any space or deep sea craft, you do have some level of moral obligation.
Whatever you feel about Elon Musk, I appreciate his distinction in approach between crewed and un-crewed craft at SpaceX. I'm paraphrasing but he has said that when something is uncrewed, you can push it to the limit over and over again to push the technology fwd. But the second you have people on board (i.e. Crew Dragon), the margin for error goes down to zero. There are unavoidable risks, but you want to be damn sure you've minimized the avoidable ones. So you can be both -- a fast moving startup, and a "safe" organization.
Obv we still know so little about the Titan, but what upsets me is that the 'uncrewed' extreme testing does not (afaik) seem to have been particularly rigorous.
If I see correctly, the base of the screen is screwed to the hull? At least there is now less than the advertised 12.7 cm of carbon fibre composite at that point.
I had a similar thought. When I saw the monitors, I thought "How are they mounted to the hull? They wouldn't screw them into it...." well, it turns out they did.
Which is a bad idea in any sense. But then, looking at the pictures, I wouldn't dive deeper than a bathtub in that thing, even if I was paid 250k to do so. It looks like a carbon fiber tube with some clued on cameras, lights, propulsion and balast. Which is far, far from being a properly designed and engineered sub. Or boat. Or piece of furniture...
Fixed, thanks. I've had a lot of thoughts about this lately. Here's another fun fact: What's wrong with this picture taken while the submersible is out of operation?
Answer: Look underneath the floor. The Altec Lansing ACS33 PowerCube 2.1 speaker system. $35 used on eBay and again from early 2000s era. -> https://i.imgur.com/fVdQ9Pc.png
Not safety critical, but so extremely indicative of how careless every element of this thing was handled before they started putting rich people in it and sending them 2.5mi underwater. It's as if they put it together by raiding my basement when I was in high school. Unbelievable.
Wow that's bad. Every new detail looks like something out of an undergrad project. It works to hack something like that together as a first prototype, but even a masters student in a decent lab with decent mentors can make something much more professional. It's like they must have actively prevented anyone with any professional experience whatsoever from seeing it.
It's likely that anyone with experience immediately walked away from the project when they saw what was going on. Who would want the reputation hit, the liabilities and the blood on their hands when things went wrong?
I think they did sue some former employee that was publicly calling out some stress detail about some window. So experienced employees did not only walk away but sounded the alarm on the way out.
It has to be hard for non-technical passagerare to judge these things, like the pre-tour disclaimer contract stating risks. I mean almost every house in CA tells you it will give you cancer.
Their customers want an adventure. Making it look like it was put together in a basement probably increases the value to their customers, compared to a 17-speaker Bose system. Of course, you would want it to look cheap but be expensive and safe.
>> You have the old guard, NASA, millions for fault analysis,
Like with most things, extremes in either direction is bad, I think there is a middle ground between using logitech controllers, and being sooo paralyzing on safety (or rather the bureaucracy of "Safety™") that not only do you never really get off the ground, but come full circle and a $0.50 o ring you failed to test in freezing conditions blows up your entire rocket even after you spent millions having human read your C code...
Just a slight correction, the o-ring not being tested in freezing conditions wasn't an oversight that destroyed the challenger; the engineers knew that the o-ring hadn't been tested at that unusually low temperature and even suspected it could fail at those temperatures.
Management at Morton Thiokol and NASA disregarded those concerns because they had already scrubbed the launch several times.
Bureaucracy is forged from the furnace of past adversities.
To further your point, I hope with venture capital funding being far more expensive due to interest rate rise, that these kinds of practices are held back a bit due to money being harder to come by.
Or because investor X wants real assurances that something like the current situation doesn’t occur.
I think a good example of a balanced company in this space, is SpaceX.
They’ve significantly reduced the cost of launching into space, have quite rapid development for their industry, yet they seem to take necessary precautions and safety measures when needed.
They aren’t afraid to test things and have them fail, and been able to not let bad PR (because some conflate successful testing with a successful launch) take them down as a company.
Maybe the “SpaceX” way to have done the sun dive would be to have a test where an automatic diver is used to send the sun down, or some first principles approach is done from the building phase. A Logitech gaming controller is surely a canary for other practices they’ve done
The "whole Silicon Valley" thing is very old guard too. Larry Ellison's first Oracle database would lose customers records randomly and reliability was left for the "future". This submarine, Amazon's Blue Origin and so many others are no exception to this rule.
I thought the latest round of NASA missions were using commodity hardware and taking the fail-early-fail-often approach (at least compared to the old days)? And in doing so have delivered, all failures included, the most bang per buck of any NASA missions?
Not saying I'd trust my life to a gamepad alone, but pragmatism is a virtue in engineering.
I’d trust it to an Xbox, PlayStation controller, they’re some of the most reliability tested input devices ever. Most people are freaking out because it’s some known glitchy budget gamepad.
From a functional safety perspective if you can't verify each step from the initial risk assessment to the final product, you can't prove the overall risk reduction and therefore don't know the overall residual hazard. From that perspective this sub would never have left dry dock (or possibly the CAD model).
In practical terms if push came to shove, an Xbox or PS controller is probably more reliable than most equivalent devices. My stock phrase when comparing normal functions with Safety Instrumented Functions is "I'd trust this with my car, but not my life."
Not quite. They're experimenting with commodity hardware to provide additional experimental functionality on missions, not to replace the core mission hardware.
For instance the mars helicopter, Ingenuity, was effectively an expendable part of the Perseverance rover mission. Even if it never took off they'd still consider the overall mission a success. That it has worked so well despite using commodity hardware has been a nice surprise and might result in NASA using more of it, but the Perseverance rover still used proper rad-hardened chips and other specialised spacecraft components.
They're also definitely not using commodity hardware and fail-fast-fail-often on crewed aircraft.
> OceanGate Inc. is a privately held U.S. company operating out of Everett, Washington, that provides crewed submersibles for industry, research and exploration. The company was founded in 2009 by Stockton Rush.[1]
"Move fast, break things" is a SV mantra. That is ok for entertainment software or social networks but when you apply this to systems where people can die like healthcare, cars, spaceships and submersibles it's a problem,
Well yes, and this enables making all the stuff that we have now like Google, YouTube or ChatGPT. If those things were developed with NASA approach to code quality, we might never have had them.
It's one thing if your favourite video host or chatbot crashes, quite another if it's a self-driven car, airplane control software or, in this case, a submarine.
I'm not saying you should hack a sub from a scrap metal and just use it as is (which seems to have happened here). But you can rapidly iterate on your project to first make it functional and then gradually improve its reliability.
What matters is not the process, but the result. For example, SpaceX are using iterative approach in their design and they now have the most reliable rocket in history.
I'm not sure of your point here. You claim that SpaceX's process was an iterative one which lead to the 'most reliable rocket in history', but state that the process does not matter.
As I see it, there are two approaches: "NASA" approach and "SpaceX" or "Silicon valley" approach. In "NASA" approach you are building the system bottom-up ensuring the reliability on each level. In "SpaceX" approach you build a barely functional system as fast as you can and then iterate to improve reliability (and other characteristics).
The top comment in this thread seems to imply that "NASA" approach is inherently better, especially for safety-critical applications. My view is that from the point of view of safety, the approach doesn't really matter and what matters is the results, i.e. reliability. At the same time from the point of view of development velocity iterative approach is clearly better.
In case of the sunk submarine, it seems that the company followed iterative, "SpaceX" approach, but they didn't actually iterate enough to make their sub seaworthy.
> My view is that from the point of view of safety, the approach doesn't really matter and what matters is the results
But you won't know the 'results' until you have a catastrophe and can do a post-mortem and find out where you went wrong. The approach does matter because one approach is 'let's do everything we can to prevent the catastrophe'. I mean, the approach is safety.
Yeah. IMHO this thing should have gone unmanned to the bottom a gajillion times, and exposed to all variety of stressors, before a paying passenger was on board. Maybe that's not possible? Maybe it was in fact prohibitively expensive? I'd love to hear why I'm wrong, but it seems wild to me that paying passengers are along for a ride after only really a handful of successful trips in this thing, ever.
SpaceX has explicitly very different approaches for crewed and uncrewed crafts. Uncrewed rockets are pushed to the limit over and over and over again, to drive innovation and thoroughly test them. As soon as a single person is on board, their margin for error goes to zero.
I'd argue this is in fact superior to NASA's approach, which rarely allows for uncrewed testing (and spectacular RUDs), leading to technological stagnation and catastrophic organizational risks. (i.e. internal resistance to doing anything about the O-ring problem on Challenger, even though they had been alerted about the potential issue).
Your first examples weren't rockets, but search engines, video hosting platforms and chatbots.
Second, it's all very well "iterating until you succeed" but that's not a good approach when live humans are involved in the iterating. "Move fast and break things" is not the same as "move fast and break people". Space X, after all, had to prove a certain level of reliability before NASA let them fly people to ISS.
If Google or YouTube has an outage people don't instantly die. Likely no one will be harmed by a bug and it can be fixed and deployed. The same isn't true for a spacecraft or submersible.
Google doesn't need belt and suspenders development, human safety systems usually do.
