LLMs in 5 years are going to be like Solidworks and similar CAD packages for the world of physical engineering - useful tools in the hands of a trained engineer, but hardly 'autonomous workers'
No one would value Solidworks at 1.5 trillion. That is the crux of the matter
An ERP system in...LISP? None of the reasons given as to why that would be a good idea are really very convincing or specific to LISP?
The main things you want for ERP are just to simplify database interactions as far as possible and to give you as many and as customizable options as possible for data visualization and curating information for a non techie user. I dont really get why LISP?
This is a very poor take and based on his education and the things he has done - he surely does not think this....he's pumping LLM based startups his/(others) companies are funding...embarassing for him intellectually but probably quite beneficial financially
Look at some of the shit John Carmack now says for God's sake. Throw enough money at anyone and they turn into an idiot
Or, maybe your understanding is wrong. He may be incentivized to say it, but you haven’t provided any reason to not believe these systems are “thinking.”
My opinions on functional languages in a practical professional setting is that they largely seem like a curiosity and a bit of an oddity when you encounter them in business software at work. I worked at a company migrating a huge amount of engineering data between two different PLM systems (in our case this was data around peristaltic pumps - 3D models, 2D designs, code, electrical schematics etc). We were using model manager for the source system and windchill for the target system and model manager uses Lisp for its automation of data management tasks and despite the fact I'd had to use Miranda (the precursor to Haskell) in my functional programming module at university (UCL) I found using Lisp for something so kind of quotidian didn't really give me the same feeling that 'this is something really special' when using it for something fairly mundane. It didn't really use any of the advantages of functional languages (and their advantages are sort of more around quite academic things anyway).
Going out of your way to use functional languages at work seems a bit like deliberately using a fountain pen and purposefully writing all business communication in a handwritten beautiful cursive style - it's kind of a cute affectation rather than actually of benefit
In the future it will be considered one of the most unusual cultural/social decisions ever, that large financial services firms are as they are in the Western world.
I have never seen a group of people so frantically doing nothing of any value.
It's one reason I am actually going to continue pursuing the 'physical engineering' side of my career rather than the software development (software 'engineering' side of it).
I studied CS at UCL years ago but worked mostly in teams with electrical, chemical, and mechanical engineers and on projects relating to various things in relation to various different industries as different as pharmaceutical and ship building. During that time I got qualified as a CAD engineer among other things I was required to by my employer(s).
Now, when I look at software jobs and the way software developers are being interviewed...I just can't. You would never ask eg a mechanical engineer in an interview 'so, you say you know SolidWorks? Ok, design me a sphere with a chamfered cylinder extruded from the centre of the top. It should be chamfered at a 46 degree angle to a plane passing through the centre of the sphere. Also, sweep a five pointed star around the sphere at the point that is midway above the plane and the top of the sphere'.
This is effectively what Leetcode tests etc are asking. It is a waste of everyone's time because they will either do it and think 'ok, that was weird' or they will refuse and think the company is insane.
There should just be accepted certifications for software development AND THEY SHOULD BE HARD. They should be very much failable and you should expect good people still to fail once or twice (like engineering or medical exams). Then just ignore the Leetcode stuff. It serves no-one.
Then all it really comes down to is frameworks etc and if someone is familiar with that particular framework or not. Just let people qualify out of this stupid algorithmic vetting process because it's utterly daft in the context of most jobs.
All very good points IMHO. Any kind of tech certification is going to need serious teeth to keep out the charlatans and outright scammers. And even that is almost certainly not going to work across borders. It's just too easy to lie, cheat and steal.
The problem I see here:
> This is effectively what Leetcode tests etc are asking. It is a waste of everyone's time because they will either do it and think 'ok, that was weird' or they will refuse and think the company is insane.
is that candidates, even excellent ones, who are in any kind of position less than fully employed and/or with f-u-money, are going to comply with this sort of mistreatment no matter how odious it seems.
I am reminded of Steve Albini's famous article on how vile the music industry's recruitment methods are.[1] (warning: ugly truth, vulgar language)
No, obviously, we're not that bad off. Yet? From what I'm told it's already effectively impossible to get a job as a junior programmer now, thanks to LLM coding assistants. Others with better pedigrees than mine have said many times here that the tech industry is eating its own seed corn with such short term thinking.
But this hiring paradigm described in TFA is making things far worse. I feel like very few people realize that tech is poised to drive off a cliff, and the few people who point this out are looked at like killjoys at best or Michael Crawford at worst.
> There should just be accepted certifications for software development AND THEY SHOULD BE HARD.
Agreed. What should these certifications test for? CompSci 101 concepts like variables, loops, good structure, maybe OOP? That's going to be useless for @koliber's Sr. Java Engineers screening problem. Heck, server jockeys like me with a smattering of shell and Perl and such under our belts could pass that.
Do we make it a dozen tests, each the equivalent of a final exam for a full semester of MIT senior-level courses? Well, now we've weeded out every possible junior starting their career; we might as well just give in to the AI assistant apocalypse right now and avoid the rush.
Do we have a certification specifically for folks who work with 3d graphics primitives and also need to be able to do matrix math in their heads?
> Then all it really comes down to is frameworks etc and if someone is familiar with that particular framework or not. Just let people qualify out of this stupid algorithmic vetting process because it's utterly daft in the context of most jobs.
I'd love to see somebody dump the HN database for the past 10 years or so and do a semantic search to count how many times this sentiment has been articulated here. If you told me it was over 10k I wouldn't be at all surprised and even if you said it was 50k I wouldn't be too skeptical.
And yet, who is actually moving the needle on correcting these systemic flaws that we all seem to recognize? Lots of folks talking about it and maybe a few companies trying, but not much progress or we wouldn't still be complaining about the same things year after year.
I have only a half-baked image of a solution in my mind, and it requires all parties to make painful sacrifices that they certainly won't do without force (ie. legislation and penalties). In another recent thread somebody mentioned that programmers are all special snowflakes who couldn't possibly be subjected to a guild or union or such. Well, we may be in the final days of the Old Times of computer programming. What happens when the tech industry has its own Quebec Bridge moment? [2] What comes next may be a mandatory Canada-style engineering culture and laws[3][4] for everybody who writes code, with prison sentences like for Practicing Medicine Without a License.
2. Some folks might say we've already had several, like Therac-25, the Snowden disclosures, the Cambridge Analytica fiasco, the vibe coder you know who dropped the Prod database because he gave Claude direct access to it, ....
I also don't know how this should work and some people's comments on this thread seem really weird to me because it seems really unlikely that someone who was eg a 'senior software developer' somewhere could be so bad that they couldn't write basic code but maybe?
In regards your comment about the Canadian idea of 'Professional Engineers' we have the same thing in the UK but it's called 'chartered engineers' and there are (more in theory than in practice) requirements around ethical and legal issues. This isn't really used in the same way I think it's used in Canada. I have worked with CEng people and I was going to go through the process (you need two sponsors you work with and you need a STEM degree that is recognised and you need 6 years of relevant experience). It's not an easy thing to get but it's not a guarantee someone is a genius or totally reliable but I haven't really seen anyone with that behaving in a way that is completely unreliable.
I wonder if it might be something to do with 'start up culture' where you've got very small teams and they're often young. You can't really tell kids to 'do a degree, get 6 years experience, qualify...then make your 'uber for milkshakes'' (or whatever they are planning to do). If you've got really young applicants you can't really expect that.
to be clear, CEng doesn't even require an exam, although everyone who would be in a position to become CEng has already done a lot of exams.
Perhaps the Leetcode thing must exist if people just refuse to qualify and want to get busy quickly. I think that is totally fine if you're making 'Uber for milkshakes' or similar but if someone is doing something that requires some responsibility then probably better to have people be vetted by accepted peers.
There are plenty of reasons that a department within a company will prefer spreadsheets. Software is not the answer to everything and also this is the same problem you get when Microsoft introduces those pesky 'Power App' developers etc or previously the Sharepoint 'web parts' etc....essentially someone who kind of feels they are the 'owner' of some information then decides to formalise the process in their own little way.
Now you have a person wasting a bit of time making their little tool but you've also got people complaining that someone's now essentially taken control of it etc. At a former employer our procurement team used spreadsheets to track basically everything but importantly everything on our PCBs - every single component and their suppliers and prices etc (various factories around China usually). This would have been a horrible thing to try and formalise in a web app just because of how often they were all changing their own conventions etc (often suppliers changing what information they gave to them, too). It would have been a wild goose chase to formalise it and more than the trouble it would be worth.
Guessing this has to do with how China tends to turn a blind eye (although increasingly less so) to copyright and more importantly trademark infringement issues of Western intellectual properties...perhaps Unity have tried to get some Chinese developers to remove things like Marvel characters, or Mickey Mouse and it's been ignored.
I think this is actually quite interesting when Unity itself has recently posted a tutorial on using its AI tools for the AI generation of a Nintendo character called Ness (main character in an old Snes game called Earthbound if you're not familiar)...seems a bit of a double standard if the complaint IS in fact in relation to trademark infringements.
There's a big problem that a lot of the people in AI have spent so much time with other PhDs in the AI space and working in these professional AI labs that they've forgotten or never even seen what 'normal' businesses look like.
I saw a podcast where the main guy behind Claude Code, Boris, just casually explained that at Anthropic where there are some software developers using sometimes multiple 6 figure values in dollars PER MONTH on their tokens for running their swarms of Claude agents...that might make you some kind of productivity guru but try going to your CFO and asking for these kinds of figures from each software developer PER MONTH...they will spit coffee out in your face and burst out laughing.
It would be the equivalent of saying 'yeah we want a new CAD and design management package that we basically prompt and it does everything for us'
'How much is it?'
'Well, upwards of 7 figures per year, per design engineer, and we have about 400 design engineers'.
'You want half a billion pounds, per year, so you can all play with your imaginary friends? No...piss off and get back to work'
No one would value Solidworks at 1.5 trillion. That is the crux of the matter
reply