So far it's about as subtle as a slap to the face. If you set ChatGPT's personality to 'straight-shooting' it starts every answer with "blunt", "tell it like it is" or "unvarnished" and permeates that across the whole reply; by its very nature, it is unable to be subtle about anything.
I have Default personality with the traits field set to straight-shooter ("Tell it like it is; don't sugar-coat responses.") I'll try the robot one, thanks.
Yeah, that's a recent addition, as far as I can tell. It became available in my account a few weeks ago. I currently use 'Robot' mode, call myself a 'Researcher', and use these custom instructions:
Answer concisely when appropriate, more extensively
when necessary. Avoid rhetorical flourishes, bonhomie,
and (above all) cliches. Take a forward-thinking view.
OK to be mildly positive and encouraging but NEVER
sycophantic or cloying. Above all, NEVER use the
phrase "You're absolutely right." Rather than "Let me
know if..." style continuations, you may list a set
of prompts to explore further topics, but only when
clearly appropriate.
Pretty happy with the results so far. Very low BS factor (although it does ignore the last part sometimes.)
This is true, but it doesn't go far enough. Google could serve you ads, like LLMs will.
But LLMs can be used to astroturf internet spaces. Which means they allow everyone the ability to serve ads and manipulate. It's no longer just limited to the company providing the original service.
The VPS is like renting office space. You don't own the space, but for the most part get to use it how you want, and all the responsibilities that come with that.
"Serverless" is like paying for a hot desk by the minute, with little control of your surroundings, but it is convenient and cheap if you only need it for an hour.
At one job I had access to some paid AWS support tier. It's basically a bunch of consultants. We needed to process a datastream of events from user actions on a website. We asked about serverless / AWS Lambda. Their answer was something like "Well yeah it'll work but don't do that. It'll cost too much money and you'll wind up rebuilding it around EC2 anyways"
Yup. If you want something to plumb some pretty low volume events, sure serverless like Lambda can be useful. Anything which would be considered high levels of compute, you are just wayyy over paying. Hell, even EC2 on spot instance is expensive compute. I do like some AWS services, but yeah they come at a premium that is just getting more and more expensive.
> Creative destruction is brutal math. The capital? Gone. Completely vaporized. But infrastructure isn’t stock certificates. Those fiber optic cables didn’t vanish when Pets.com did.
I read quotes like this and reminded that it is common that people forget money is just a competitive resource we use to outbid each other for _real_ things. Money moves around, it isn't lost or "Completely vaporized", someone receives it at the other side of the transaction. It is still in circulation, it can still be used to outbid people for real things, just by different people.
Also, pets.com still exists, it just forwards to petsmart.com.
Money can be "lost", and "created". In fact, it regularly is by commercial banks; this is the cornerstone of the modern economy. The bank of England wrote a pretty accessible document on money being created and destroyed [1]. For a slightly deeper dive (but equally accessible) check out "Can’t We Just Print More Money?" by Rupal Patel, et al. [2] which describes the different kinds of money.
The 2014 doc was a pretty wild read for me when it came out - it changed my perspective quite a bit.
> money is just a competitive resource we use to outbid each other for _real_ things
That's true, but the thing that's lost is the economic/productive capacity that the money was spent on, that could have been used for other (better) purposes.
For example, if I raise $100mn in a frothy market, and spend it on employing 100 Engineers on $1mn/yr salaries for 1 year before ultimately going bankrupt, it's true that the money doesn't disappear, as it was simply transferred from the VCs to the Engineers, but what's spent/consumed is the Engineers' time. Society can never get those 100 person-years back, and the VCs have to write their capital investment to 0.
The other comments are separately true - money is created by bank borrowing and destroyed by loans being repaid or going bad. Periods of speculation often result in increasing leverage (e.g. borrowing to buy stocks/houses), which does result in the destruction of money when it unwinds (as well as damage to bank's balance sheets, which can become problematic when it happens at a large enough scale - see 2008).
> I read quotes like this and reminded that it is common that people forget money is just a competitive resource we use to outbid each other for _real_ things.
But money is an abstraction of wealth, and wealth absolutely can be destroyed, in multiple ways:
1. It can be physically destroyed - if I break a window, that's wealth that is destroyed. That window now needs to be replaced, which costs materials and labor, which could've gone to building something new instead.
2. It can be spent on things that end up not used. If five years from now, those millions of GPUs are no longer in use, we created them for nothing instead of creating more of something people would use.
3. Wealth can be spent on the less important things, rather than the more important things. This is not exactly wealth being destroyed, just built more slowly, because instead of building lots of new wealth (via innovation, say) we're creating less valuable things.
I don't think any of the above are relevant to AI, btw.
What I find so weird always is that well value "vaporized" when prices went down. But what is the term when prices went up? As surely that wealth also respectively came from same place it went when it went down...
"capital" in the economic sense isn't money at all - it's tools, infrastructure, knowledge; from this point of view, you're correct. A bubble bursting on the stock market doesn't destroy capital.
On the other hand, the monetary value of the stock market (and other assets) going up and down does create or destroy "money". From a financial point of view, it's not a zero sum game.
This is incorrect. https://en.wikipedia.org/wiki/Money_creation . A bubble bursting is the inverse process of money creation. Money isn't paper bills or metal coins. It is mostly numbers in a computer.
Same, early 2010s IMO seems to be the point where the industry really started to shift. There are some good cars from before this time, but keep them running past the 2030s will be a challenge.
1980s widespread adoption of electronic fuel injection - this is generally a good thing, cars become more complex but run better more of the time
1990s widespread adoption of more advanced emissions control systems - for reliability i'd say this is only a backwards step - none of these systems are required to propel the car down the road but many of them can stop a car from driving. They are additional complexity, weight and cost for limited functional benefit (in this generation, fuel economy improvements were fairly small compared to the leap from carbs to EFI in the previous gen).
2000s widespread adoption of on-car networks, the emissions diagnostics technology introduced in the previous decade was now no longer the primary use of on-car networks. Now your car stereo knew how to increase its volume as your road speed increased etc. screens became larger and colourful. Onboard software (typically bug ridden) became a security risk.
2010s widespread adoption of telematics maybe? That was more mid-late 2010s though
According to the industry people stopped doing maintenance and were more likely to trade in their vehicle for a new one. So they stopped optimizing for that segment of the audience and started making disposable cars.
Truthfully, industry watched the government bail out the banks, require next to nothing in return, and demanded no prosecutions for illegal behavior. The writing was pretty much on the wall. Industry realized it no longer needed happy customers.
>> According to the industry people stopped doing maintenance
Nah the motoring industry has been saying that forever. Just to back that up, I inherited a ton of the "Car Mechanics" magazine here in the UK. Just for fun I've just pulled a random 1960s one - November 1965, 2 shillings it cost. Flicking through, firstly i'm struck by just how many adverts there are, there's 1.5 pages of advert per 0.5 page of content and the adverts are commonly absolute tat like antifreeze additives that surely do nothing useful. Anyway, by page 35 we've found our first nobody maintains anything anymore story: "Transport Tests", top right hand side - a column railing against "defective lorries on the road", "49% of trucks stopped were defective" - i knew no matter which magazine i pulled, there'd be something in there decrying a lack of maintenance these days (in 1965...).
>> So they stopped optimizing for that segment
Again, i don't think you're accurate here. There's nothing a mechanic loves more than to gripe about the engineers who foolishly designed a car to be harder to service. It's a time-tested complaint of the mechanic. Today with all the tight packaging of various systems i think there's often a point to be made about ease of repairability but even when engine bays were gaping empty holes in the 50s and 60s that you could literally stand in (in some cases) with plenty of access space while you worked, there were models of cars derided as "hard to work on" because of lack of optimisation for maintenance. Packaging is hard. If the decision is between optimising for sale (aesthetics, packaging etc) or for maintenance, the design engineer is going to lose that battle in the drawing room.
I think market forces have changed how manufacturers view servicing over time though. If you're doing fleet sales, and your product requires an oil change every 8k miles and the other manufacturer is every 20k miles, then the company purchaser who cares not a jot about mechanical sympathy and a whole lot about a bonus for saving the firm money, applies a pressure to the market to reduce maintenance costs over the first 3-4 years that a vehicle is leased for. And so today i own a van with an absolutely absurd oil change interval of 25k miles.
You're approaching this as if it would be a binary across an entire industry. That's obviously not the correct level of analysis. The question is, "are more and more manufacturers doing this?" You've made no case to this point.
> And so today i own a van with an absolutely absurd oil change interval of 25k miles.
Would you share the Year, Make and Model please? I want to see this and I want to see precisely which oil products they recommend. The part you may have missed is that specialty oils have come a long way in the last 2 decades.
Software updates and data collection. Eg, my mums Toyota Corolla 2018 already has a disabled infotainment button because of dropped support. If it was a 2019 it would have been eligible for an update, but not for 2018.
Lots of cars from the same period are collecting and sharing data to various different companies from weather to insurance.
Personally I don’t want monitoring or software updates, and definitely don’t want any cloud dependencies.
I'd say Tesla built a futuristic computer on wheels, with huge screens, always-on internet connectivity, smarter remote features than most other cars, etc. The car itself was exotic enough by being an EV and that drew attention to these other features too. Everyone else started to emulate them for better or worse.
For now these things are modular because it was the cheapest way to build them. If manufacturers get over the hurdle of cost and find a way to have everything more vertically integrated (think Apple) then we'll lose all access to tinker with the hardware which might be a couple of black-box chips, or the software.
This is probably what Apple was trying to sell as a smart car to car manufacturers. They might have dropped those plans to focus in CarPlay and have the phone be that "smart car". Hopefully some brands go the other way and make a dumb car where the brain is entirely the phone but that's handing out a lot of their agency to the phone manufacturer.
> 2010s widespread adoption of telematics maybe? That was more mid-late 2010s though
I haven't done it yet, but maybe looking into the EU mandatory regulations would make sense. eCall, for instance (a feature that will call for help if you crashed by contacting an operator), was made mandatory in new cars in 2018. The initiative gained traction at around 2013.
Hmmm, but I think that was already happening in the early 2000s. It definitely accelerated in the late 2000s.
I worked for a deutsche telekom subsidiary in 2004 and they had a brand new BMW e65 7 Series all liveried up in corporate branding, advertising internet connectivity on the move. At the time i thought that was the ultimate car but i never managed to borrow it...
On an adjacent theme there has been a large debate in Sweden if a car that has mandated automatic eCall over 2G or 3G is faulty and thus not roadworthy if the 2G and 3G networks are turned off. The eCall feature was introduced broadly in 2018. It still uses modem sounds over a voice phone call to relay car emergency status and position. New solutions for packet only networks (4g and 5G) have been standardised but are not retrofitted to older cars.
I had thought most of europe was keeping 2G service online. Too many deployed devices that are low traffic but expensive to replace modules on. Even if they had a 3g module, they'll fall back to 2g, so there's no sense in using spectrum for 2g and spectrum for 3g.
The first way this can go is that the people with a stake in selling cars (dealers, OEMs, etc) get their lobbyists leaning on the politicians and get it rammed through.
The second is that the issue gets delayed long enough that the number of older vehicles it applies to goes down out of attrition and it's not worth fighting over.
Considering the cultural disposition of the nordics when it comes to matters like this I expect the first option to be chosen. There will enough people hand wringing about safety and whatnot to provide the political lubricant to ensure the first outcome.
In Australia, AWS dev position in 2021 had a take home. Microsoft contract position before that didn’t but they were desperate to fill seats on a poorly executed gov contract.
Claude is decent for sure, but if you are using these models for 'smarts', that is a whole separate problem. I also think honestly people are sleeping on Mistral's medium 3 and devstral medium. I know it isn't 'smart' either (none of them are), but for mundane tasks need valid code output, it is extremely good for the price.
As a sounding board for things you are already well familiar with, I agree, and have experienced the same, and that can be useful. It's also a much better experience than say using Google to do the same, or just a rubber ducky.
The NLP these models can do is definitely impressive, but they aren't 'thinking'. I find myself easily falling into the habit of filtering a lot of what the model returns and picking out the good parts which is useful and relatively easy for subjects I know well. But for a topic that I am not as familiar with, that filtering (identifying and dismissing) I do is much less finessed, and a lot of care needs to be taken to not just accept what is being presented. You can still interrogate each idea presented by the LLM to ensure you aren't being led astray, and that is still useful for discovering things, like traditional search, but once you mix agents into this, things can go off the rails far too quickly than I am comfortable with.
Spot on. The issue I think a lot of devs are grappling with is the non deterministic nature of LLMs. We can protect against SQL injection and prove that it will block those attacks. With LLMs, you just can’t do that.
It's not the non-determinism that's a problem by itself - it's that the system is intended to be general, and you can't even enumerate ways it can be made to do something you don't want it to do, much less restrict it without compromising the features you want.
Or, put in a different way, it's the case where you want your users to be able to execute arbitrary SQL against your database, a case where that's a core feature - except, you also want it to magically not execute SQL that you or the users will, in the future, think shouldn't have been executed.
> it's that the system is intended to be general, and you can't even enumerate ways it can be made to do something you don't want it to do, much less restrict it without
Very true, and worse the act of prompting gives the illusion of control, to restrict/reduce the scope of functionality, even empirically showing the functional changes you wanted in limited test cases. The sooner this can be widely accepted and understood well the better for the industry.
I’m not a fan of k8s, but “serverless” imo is not a good trade either, especially AWS lambda. ECS with ALB is probably an ok middle, but still expensive, even with heavy use of Spot instances. Cloud isn’t going to be getting cheaper..
Aren't we exaggerating the whole Kubernetes is difficult thing a bit? Unix is also more complex than DOS, still glad we didn't get stuck with the latter bc it was "good enough"
reply