lordnacho
ProphetX is one of the fastest growing prediction markets in America. Fully CFTC regulated and particularly big in the sports area, we are looking to fill a number of roles across engineering and product.
Among others, we are looking for:
Engineering Manager
Staff Backend
Senior Backend
Staff DevOps (note this hasn't appeared on the job site yet, check in a day or so). This one should be open for EU timezones as well.
It's not terribly important which stack you are in, but we are building some low latency, high throughput distributed systems.
Please mention Hacker News so I can show everyone the wonderful quality of candidates available on this board.
Why should I learn Stoke's Theorem? Why should I learn Nietzsche? Questions you can't answer until you learn them, and questions that compete with "how will I find a job" and "how will I find a girlfriend".
The problem is, if you don't learn how to answer such questions, you will not know that such questions exist. You will find a job and a partner, and think you never needed to contemplate any of the things you didn't think about.
But I also care about the complexity classes of the solution. I prod it about the time and space complexity, and about cache locality.
I do a lot of black box testing: take the solution, see if it does what I expected, then bake in those expectations in a document.
I find that if I do those things, it produces better code than I could do in the same time.
I also think it's more or less the same as how I used to work, just more thorough and with fewer time-consuming, trivial bugs.
I actually found code exactly like that would break a test on a certain date, using an agent.
It's really good at flagging things like this, which tend to be leftovers from a human process that nobody has had the time to clean.
You can then build a statistical arbitrage alpha based on the difference between an ordinary amount of selling (conditioned on a bunch of stuff) and a lack of selling/accelerated selling.
If you're building a home lab, you might just give up because you are tired of searching for Cisco router syntax.
LLM can be the little bit of lube you need to understand things at a higher level, because the tedious things are not as bad as they used to be.
You then end up actually getting a few things in your lab to work correctly, get a confidence boost, and follow up with some searches about how routing actually works and why the syntax is the way it is.
If the parrot randomly yaps out a solution to an open math problem, it's a useful parrot.
The fact that it sometimes says things that aren't useful, well, old intelligence does that all the time.
If you've used AI recently, you can't avoid having found it useful. Just general day-to-day intelligence like finding things in your email, writing code for you, that kind of stuff, it does well. People have some weird hangups about it being stochastic, when it doesn't really matter.
He is right that the AI giants would like a bit of regulatory capture.
If it were me, I would start by getting a summary of what problems the existing solution solves. I would also make sure that summary had the non-functional requirement. Maybe write a test suite that opens the page and does all the stuff.
Then ask the LLM to implement in another language, with those solutions in mind, but first coming up with reasons why the target language might not be as good, and where it might have useful features that the original language didn't. Use the test suite to check if things are working.
I honestly don't think you would land far away. You would have to answer a few questions along the way, but it would mostly be plain sailing.
I would expect to be able to do this with very little human attention, whereas a language rewrite two years ago might take a whole month, not including learning the new language.
There's only 500 numbers it could be, assuming someone knows those other things about you.
In any case, there are alternative systems for authorisation.
What was it good for? Well, often it was prompting a person in a low-cost location what code to write. He would pick it up in his timezone and implement it overnight, and present findings/questions as a response.
The problem is, coding has made implementation super cheap. It is now so cheap you have to compare it to the cost of dealing with someone in another timezone, who does not have the same cultural context on the problem. They also can't ask you clarifying questions while you're asleep. AI wins against the offshore guy, on both money and context.
Who does it help? Amazingly, the experienced onshore dev who thought he was going to be unemployed past 40. This person is now able to implement several things concurrently, has experience in making technical decisions, and has local cultural context.
The only thing I can think of is the reduction of truffles as an ingredient, which apparently happened around WWI. We eat a lot fewer of them than before.
> Second, I'm really curious what's your "adapt" path, because I suspect you're in denial too. Accept that, before long, an LLM will not need you to write good prompts, tell it to "keep going", or say "no, that's wrong". What's your value then? Does your employer keep you on the payroll as charity? For old times' sake? If your value is just supervising AI, then your value is declining rapidly.
Your employer cannot talk to all the LLM sessions at the same time. He also can't write a prompt that will do it for him.
> People in unskilled professions don't have choice when automation shows up and takes their jobs. Software engineering may go down in history as one of the few highly skilled professions that didn't just keel over and die, but enthusiastically embraced its own demise.
It's changed several times over the decades. How many people are working on a mainframe nowadays? How many people are writing assembly? Look at how many languages have come and gone too. LLM is a big change, but that doesn't mean software people will be gone.
In fact I'm more or less at the point where I don't even want the model to get better, I care more that it outputs faster.
A year ago, I asked ChatGPT to design me a 3D model for a wallet, the kind you can stick to the back of a phone with a magnet. It screwed it up over and over, never got to anything usable.
Last week I tried again, just explaining in words what to do, and it made what I had in my mind.
For websites, it's over. Any kind of brochure/simple restaurant booking site, that was easy years ago and trivial now. SaaS is going next, of course there's a lot of types of SaaS but your average "web implementation of commonly needed service" is now very easily copied with AI.
Anything like a task list or information management system with no particular constraints in terms of performance or scale, you can just make an AI customize for you.
Now, it was already like that to an extent. To be crude, the job was "decide what needs to be made, and make it". People would be on various splits between the two, typically on the "make it" end if they were junior.
Initially, I thought about what effect this would have on hiring. Perhaps we would get a lot more people who were not so great at implementation, since we are hiring for judgement?
Let's consider Berkson's paradox.
Let's say that there's a correlation between implementation skill and engineering judgement, eg r=0.5, a rugby-ball shaped cloud that points up and to the right. The basis for this being that you maybe learn good judgement from implementing things, something like that.
Before AI, selection on the sum scores (judgement + implementation) would give you a set of engineers whose correlation was deflated from the underlying correlation. It could even go negative (in fact it would if the underlying correlation was zero, but we assumed it was positive). This is the original Berkson's paradox, selection makes the correlation coefficient lower. There would be less of a connection between someone's (who had been hired) judgement and implementation skill than 0.5, it would tend downwards, possibly landing at zero.
Now, let us imagine that AI means the implementation side becomes less important. We will veer towards hiring people based on engineering judgement, rather than coding skill. Suppose we add a parameter on (judgement + lambda * implementation).
What actually happens is, if lambda tends towards zero, we are selecting on a line that is only judgement. But that means actually, the deflating effect of the selection is lowered, and we get a correlation between our variables that is closer to the underlying set (it won't necessarily be all the way back, since we restrict the range).
So then actually, somehow, people we hire who are good at judgement will also be good at implementation. Somewhat unexpected.
Whatever it is being asked, it is great at throwing out new suggestions. This MRI idea sounds like another one of them.
It works well because humans are crap at checklists. For more or less anything you can think of, you'll come up with maybe 4 items, and the AI will add a few more. They may not be relevant, but the AI is very thorough.
Imagine you are evaluating what the machine should do when it is improving itself. It does a bunch of work and returns with "I supervaluated the liminal overdecomposition from the previous homological calibulation pass. It shows us that subtransitory mulutination will underspecify the tensor of stermullification. Where do you want to go from here?"
It will be like when you are reading a Wikipedia about a topic you don't understand. You follow the links, and you get more questions with more links. Your whole day is taken up following links, to the point where you forgot the original question.
Except this time, all the words come from the AI's work. You can't refer to an external authority who has already been there and can tell you what to do.
The AI needs you to tell it whether it is more intelligent than it was before, but you don't know, because you can't follow its reasoning any more. It's like an ordinary person trying to hire a math professor, there's just no way to do it.
But whereas a human math prof can evaluate another one, a machine intelligence can't evaluate another one, by construction. Because it's still usefulness to humans that is the evaluation criterion.
Not sure how I'd share it. I suppose I could order a copy for you.
I've gotten AI to generate an entire book, with the plot points kept coherent (this character knows that when this chapter starts, that kind of thing). It didn't impress my wife, but it was a book.
I would think just a coarse sorting would do, paired with a free meal. Every couple of weeks, you go for a random date with a somewhat well paired person, and see what happens. Maybe you don't meet the love of your life, but you will meet someone roughly your age living in the same society, a perspective worth having regardless of whether you're compatible.
However, if you know anyone from Singapore, you must think this is a rather small plaster on a gaping wound.
Singapore is like a permanent examination hall. Everyone is competing to do well on exams, and then they try to get the best internships, and then the best jobs. It's not exactly conducive to starting a family. You'll lose a step in the rat race, and then your kid will need to join this rat race.
However I hit a wall with the game that has a relation to how AI has upended the programming world.
Basically, I got bored, and it might be because of the automation.
In the early game, you are running around with an axe, manually putting down inserters and assemblers, manually crafting items. You eventually make some assembly lines to automatically build stuff, and the outputs then go into more and more complex recipes for which you need to organize the flow of resources.
Along the way, you gain the ability to build trains to move goods from faraway places. Along with a number of other mechanics that are introduced, this creates a bit of novelty, because you discover some slightly different requirements and you get to exercise your brain in how to connect things together.
But then at some point, you learn how to make robots. Robots can place factory components as well as just moving goods around. You make a blueprint, plonk it down, and then the robots get all the necessary pieces, and place them down for you. This completely alters the game, because a blueprint can have a huge number of pieces while still being modular. There are people with libraries of square tiles encompassing every aspect of the game, that can simply be snapped together like legos.
I got to a point where I was assembling libraries and caring more about building a good set of components than actually wiring things together into a factory. I got an auto-placement mod, which I figured was a bit like predictive text in an IDE, it saves you from having to put down inserters and turn them, that kind of micro thing that can mess up your factory if you screw it up, but is actually tedious.
Armed with my new library of everything and my auto-placement mod, I destroyed my enjoyment of the game, somehow.
I had played the game enough times to understand what was needed, and so I got to a place where I just needed to place down a mall (a special kind of factory that mainly makes components) along with the bus (assembly lines of outputs), and things kinda couldn't go wrong. I didn't do any more of the typical early-game fiddling with making ad-hoc designs, because I knew what would work, and I didn't want to do the micro. I just satisfied myself that a few belts and rail lines would be enough to glue my libraries together, and that was that.
What does this have to do with AI? Well, I hear a lot of people's complaints with AI is that they were no longer actually coding. Without the need to type out anything, the micro is gone. People were already glueing libraries together, but at least there was some delight in writing the little bits of knitting. Now, you can just execute your architecture with a prompt, and the LLM will just do it all for you. There's still a necessity of checking what comes out, but the character of the work has changed.
When you see an attractive woman, do you have a choice about whether you think she's attractive?
I would say that is so instinctual as to not be a choice, even though different people have different opinions about who is attractive. And yet we know that men in particular go to great lengths to be with attractive women, some to the point where it becomes an obsession.
What about food? Do you not like things that have a bit of fat, sugar, and maybe some crunch to it? There are clearly people who, despite having all the necessary information, will eat too much, to the point where they become ill from it. How is that a choice?
So, what does it mean to choose something with full agency?