If the offshore company provides me a Rust crate that compiles, that is already a lot of guarantee. Now that does not solve the logic issues and you still need testing.
But testing in Python is so easy to abuse as LLM. It will create mocks upon mocks of classes and dynamically patch functions to get things going. Its hell to review.
> Well, I am on the provocative side that as AI tooling matures current programming languages will slowly become irrelevant.
I have the opposite opinion. As LLM become ubiquitous and code generation becomes cheap, the choice of language becomes more important.
The problem with LLM for me is that it is now possible to write anything using only assembly. While technically possible, who can possibly read and understand the mountain of code that it is going to generate?
I use LLM at work in Python. It can, and will, easily use hacks upon hacks to get around things.
Thus I maintain that as code generation is cheap, it is more important to constraint that code generation.
All of this assume that you care even a tiny bit about what is happening in your code. If you don't, I suppose you can keep banging the LLM to fix that binary blob for you.
> The problem with LLM for me is that it is now possible to write anything using only assembly. While technically possible, who can possibly read and understand the mountain of code that it is going to generate?
As a very practical problem the assembly would consume the context window like no other. And another is having some static guardrails; sometimes LLMs make mistakes, and without guard rails it debugging some of them becomes quite a big workload.
So to keep things efficient, an LLM would first need to create its own programming language. I think we'll actually see some proposals for a token-effective language that has good abstraction abilities for this exact use.
> As LLM become ubiquitous and code generation becomes cheap, the choice of language becomes more important.
I think, changes to languages/tooling to accomodate Agentic loops will become important.
> All of this assume that you care even a tiny bit about what is happening in your code. If you don't...
I mean, as software engineers, we most certainly do. I suspect there'll be a new class of "developers" who will have their own way of making software, dealing with bugs, building debugging tools that suit their SDLC etc. LLMs will be to software development what Relativity was to Astrophysics, imo: A fundamental & permanent shift.
We should treat phones on kids the same we treat alchohol. "What the fuck, is that a phone? Give me that!" The only other solution involves evaporating our privacy. Fuk 'em kids. I guess they don't get to use phones, we survived, why can't they?
In fact, it is probably better for them to "struggle" and figure out by themselves how to find a way to circumvent it. Make them think instead of having thoughts feed into them.
#### The Paradigm Shift Brought by Palantir: Ontology as an Operational Layer
The *"Ontology"* strategy by Palantir, explained in this book, is a paradigm shift that fundamentally breaks this deep-rooted disease of silos.
In the context of knowledge engineering and the semantic web, the widely cited academic definition of "ontology" is an "explicit specification of a conceptualization" by Gruber (1993).
Furthermore, Studer et al. (1998) expanded on this, proposing the definition of a "formal, explicit specification of a shared conceptualization."
This transition from "data just for viewing" to "data that directly drives the business" is the key to true digital transformation in the AI era.
[end]
Just gave me brain damage. Please for the love of god just go straight to the point. Just give me the prompts that wrote all of this.
> But yeah, for web stuff? Just use TypeScript. Life's too short to fight the borrow checker over a blog.4
I am not sure about TypeScript. I think having static typing is just too good of an insurance against stupid bug and for your own sanity. I think for web purposes, especially with LLM around, you probably should just use Go. You don't have to like it, but there's enough training dataset for your CRUD application. So all you really need to do is to be able to read it.
> An interesting thought experiment is whether it's possible that AI tools could write a novel that's better than War and Peace. A quick google shows a lot of (poorly written) articles about how "AI is just a machine, so it can never be creative," which strikes me as a weak argument way too focused on a physical detail instead of the result. War and Peace and/or other great novels are certainly in the training set of some or all models, and there is some real consensus about which ones are great, not just random subjective opinions.
I am sure it could but then what is the point? Consider this, lets assume that someone did manage to use LLM to produce a very well written novel. Would you rather have the novel that the LLM generated (the output), or the prompts and process that lead to that novel?
The moment I know how its made, the exact prompts and process, I can then have an infinite number of said great novels in 1000 different variations. To me this makes the output way, way less valuable compared to the input. If great novels are cheap to produce, they are no longer novel and becomes the norm, expectation rises and we will be looking for something new.
I'm inclined to believe that the difference that makes the upper bound of human writing (or creativity) higher than that of an LLM comes from having experiences in the real world. When someone is "inspired" by others' work or is otherwise deriving ideas from them, they inevitably and unavoidably insert their own biases and experiences into their own work, i.e. they also derive from real-world processes. An LLM, however, is derived directly and entirely from others' work, and cannot be influenced by the real world, only a projection of it.
> Would you rather have the novel that the LLM generated (the output), or the prompts and process that lead to that novel?
The "process", in many cases, is not necessarily preferable to the novel. Because an important part of the creative process is real-world experiences (as described above), and the real world is often unpleasant, hard, and complex, I'd often prefer a novel over the source material. Reading Animal Farm is much less unpleasant than being caught in the Spanish Civil War, for example.
I also think it's a matter of time before we start constructing virtual worlds in which we train AI. Meaning, representations of simulated world-like events, scenarios, scenery, even physics. This will begin with heavy HF, but will move to both synthetic content creation and curation over time.
People will do this because it's interesting and because there's potential to capitalize on the result.
I thought of this in jest, but I now see this as an eventuality.
> People will do this because it's interesting and because there's potential to capitalize on the result.
I don't know why anyone admits to thinking this. For one, there's nothing stopping you from making movies or writing stories now. You're not suddenly going to develop creativity or interesting ideas using LLMs, either.
Also, think it through. If everyone can yell at computer until movie fall out, there will be millions of them and nobody will pay for anything.
> The "process", in many cases, is not necessarily preferable to the novel. Because an important part of the creative process is real-world experiences (as described above), and the real world is often unpleasant, hard, and complex, I'd often prefer a novel over the source material. Reading Animal Farm is much less unpleasant than being caught in the Spanish Civil War, for example.
I think you misunderstood what I meant by "prompts and process that lead to that novel". I am talking about the process that the "author" used to generate that novel output. I am more interested in the technique that they use, and the moment that technique is known. Then, I can produce billions of War And Peace.
I suppose the argument is that, the moment there's an LLM that can produce a unique and interesting novels, what stops it from generating another billion similarly interesting novels?
This so fundamentally misunderstands (1) the point of writing a novel and (2) what makes a novel interesting.
A novel isn't just a buncha words slapped together, bing bam slop boom, done.
What makes a novel interesting is the author and the author's choices, like all art. It's the closest you can get to experiencing what it's like to be someone else. You can't generate that, it's specific to a person.
The GP assumes that an LLM is able to write such novel. So I was working from there. My thesis is that even IF LLMs are able to produce "novelty", it will become the norm and we will simply demand even more exotic novelty.
> An interesting thought experiment is whether it's possible that AI tools could write a novel that's better than War and Peace. A quick google shows a lot of (poorly written) articles about how "AI is just a machine, so it can never be creative," which strikes me as a weak argument way too focused on a physical detail instead of the result. War and Peace and/or other great novels are certainly in the training set of some or all models, and there is some real consensus about which ones are great, not just random subjective opinions.
It can have anything you like in a training set, you still can't build specific human experiences.
I haven't read War & Peace -- I don't have the patience for Russian literature -- but a much more accessible example is the Vorkosigan series by Lois Bujold. She uses a lot of Tolstoy lol.
While you can read them as fun military scifi, that's not why the series is so good and so famous. In her books, humanity invented two critical things: wormhole FTL travel and uterine replicators.
A lot of the series is exploring how people actually would use and abuse those two things. And then on another layer the books are about her thoughts on parenting, marriage, power, inheritance, and so on.
Good art isn't about accepting someone's opinion that it's good art. Good art impacts you. I think about things differently after those books.
You cannot write a good novel using the algorithmic mean of a lot of different stories.
This is one of my major concerns about people trying to use these tools for 'efficiency'. The only plausible value in somebody writing a huge report and somebody else reading it is information transfer. LLM's are notoriously bad at this. The noise to signal ratio is unacceptably high, and you will be worse off reading the summary than if you skimmed the first and last pages. In fact, you will be worse off than if you did nothing at all.
Using AI to output noise and learn nothing at breakneck speeds is worse than simply looking out the window, because you now have a false sense of security about your understanding of the material.
Relatedly, I think people get the sense that 'getting better at prompting' is purely a one-way issue of training the robot to give better outputs. But you are also training yourself to only ask the sorts of questions that it can answer well. Those questions that it will no longer occur to you to ask (not just of the robot, but of yourself) might be the most pertinent ones!
Yep. The other way it can have net no impact is if it saves thousand of hours of report drafting and reading but misses the one salient fact buried in the observations that could actually save the company money. Whilst completely nailing the fluff.
> LLM's are notoriously bad at this. The noise to signal ratio is unacceptably high
I could go either way on the future of this, but if you take the argument that we're still early days, this may not hold. They're notoriously bad at this so far.
We could still be in the PC DOS 3.X era in this timeline. Wait until we hit the Windows 3.1, or 95 equivalent. Personally, I have seen shocking improvements in the past 3 months with the latest models.
Personally I strongly doubt it. Since the nature of LLM's does not allow them semantic content or context, I believe it is inherently a tool unsuited for this task. As far as I can tell, it's a limitation of the technology itself, not of the amount of power behind it.
Either way, being able to generate or compress loads of text very quickly with no understanding of the contents simply is not the bottleneck of information transfer between human beings.
Yeah, definitely more skeptical for communication pipelines.
But for coding, the latest models are able to read my codebase for context, understand my question, and implement a solution with nuance, using existing structures and paradigms. It hasn't missed since January.
One of them even said: "As an embedded engineer, you will appreciate that ...". I had never told it that was my title, it is nowhere in my soul.md or codebase. It just inferred that I, the user, was one. Based on the arm toolchain and code.
It was a bit creepy, tbh. They can definitely infer context to some degree.
> We could still be in the PC DOS 3.X era in this timeline. Wait until we hit the Windows 3.1, or 95 equivalent. Personally, I have seen shocking improvements in the past 3 months with the latest models.
While we're speculating, here's mine: we're in the Windows 7 phase of AI.
IOW, everything from this point on might be better tech, but is going to be worse in practice.
Context size helps some things but generally speaking, it just slows everything down. Instead of huge contexts, what we need is actual reasoning.
I predict that in the next two to five years we're going to see a breakthrough in AI that doesn't involve LLMs but makes them 10x more effective at reasoning and completely eliminates the hallucination problem.
We currently have "high thinking" models that double and triple-check their own output and we call that "reasoning" but that's not really what it's doing. It's just passing its own output through itself a few times and hoping that it catches mistakes. It kind of works, but it's very slow and takes a lot more resources.
What we need instead is a reasoning model that can be called upon to perform logic-based tests on LLM output or even better, before the output is generated (if that's even possible—not sure if it is).
My guess is that it'll end up something like a "logic-trained" model instead of a "shitloads of raw data trained" model. Imagine a couple terabytes of truth statements like, "rabbits are mammals" and "mammals have mammary glands." Then, whenever the LLM wants to generate output suggesting someone put rocks on pizza, it fails the internal truth check, "rocks are not edible by humans" or even better, "rocks are not suitable as a pizza topping" which it had placed into the training data set as a result of regression testing.
Over time, such a "logic model" would grow and grow—just like a human mind—until it did a pretty good job at reasoning.
> I would like to see the day when the context size is in gigabytes or tens of billions of tokens, not RAG or whatever, actual context.
Might not make a difference. I believe we are already at the point of negative returns - doubling context from 800k tokens to 1600k tokens loses a larger percentage of context than halving it from 800k tokens to 400k tokens.
There's many things that used to be called AI, but as their shortcomings became known we started dropping them from the AI bucket and referring to them by a more specific name: expert systems, machine learning, etc. Decades later plenty of people never learned this and those things don't pop into mind with "AI" so LLMs were able to take over the term.
Hehe, yeah there's some terms that just are linguistically unintuitive.
"Skill floor" is another one. People generally interpret that one as "must be at least this tall to ride", but it actually means "amount of effort that translates to result". Something that has a high skill floor (if you write "high floor of skill" it makes more sense) means that with very little input you can gain a lot of result. Whereas a low skill floor means something behaves more linearly, where very little input only gains very little result.
Even though its just the antonym, "skill ceiling" is much more intuitive in that regard.
Are you sure about skill floor? I've only ever heard it used to describe the skill required to get into something, and skill ceiling describes the highest level of mastery. I've never heard your interpretation, and it doesn't make sense to me.
Yes, I am very sure. And it isn't that difficult to understand, it is skill input graphed against effectiveness output. A higher floor just means that with 1 skill, you are guaranteed at least X (say, 20) effectiveness output.
The confusion comes from people using "skill floor" for "learning curve" instead of "effectiveness".
But this is a thing where definitions have shifted over time. Like jealousy. People use "jealousy" when they really mean "envy", but correcting someone on it will usually just get you scorn and ridicule, because like I mentioned, language is fluid.
If the skill floor is high and therefore "effectiveness" is the same for a wide range of skill levels, isn't that the same as having a high barrier to entry? It seems that any activity or game where it takes a lot of skill before you can differentiate yourself from other players would be described that way.
It reminds me of that Apple ad where a guy just rocks up to a meeting completely unprepared and spits out an AI summary to all his coworkers. Great job Apple, thanks for proving Graeber right all along.
> Those questions that it will no longer occur to you to ask (not just of the robot, but of yourself) might be the most pertinent ones!
That is true, but then again also with google. You could see why some people want to go back to the "read the book" era where you didn't have google to query anything and had to make the real questions.
One thing AI should eliminate is the "proof of work" reports. Sometimes the long report is not meant to be read, but used as proof somebody has thoroughly thought through various things (captured by, for instance, required sections).
When AI is doing that, it loses all value as a proof of work (just as it does for a school report).
My AI writes for your AI to read is low value. But there is probably still some value in "My AI takes these notes and makes them into a concise readable doc".
> Using AI to output noise and learn nothing at breakneck speeds is worse than simply looking out the window, because you now have a false sense of security about your understanding of the material.
i may put this into my email signature with your permission, this is a whip-smart sentence.
and it is true. i used AI to "curate information" for me when i was heads-down deep in learning mode, about sound and music.
there was enough all-important info being omitted, i soon realized i was developing a textbook case of superficial, incomplete knowledge.
i stopped using AI and did it all over again through books and learning by doing. in retrospect, i'm glad to have had that experience because it taught me something about knowledge and learning.
mostly that something boils down to RTFM. a good manual or technical book written by an expert doesn't have a lot of fluff. what exactly are you expecting the AI to do? zip the rar file? it will do something, it might look great, lossless compression it will be not.
P.S. not a prompt skill issue. i was up to date on cutting edge prompting techniques and using multiple frontier models. i was developing an app using local models and audio analysis AI-powered libraries. in other words i was up to my neck immersed in AI.
after i grokked as much as i could, given my limited math knowledge, of the underlying tech from reading the theory, i realized the skill issue invectives don't hold water. if things break exactly in the way they're expected to break as per their design, it's a little too much on the nose. even appealing to your impostor syndrome won't work.
P.P.S. it's interesting how a lot of the slogans of the AI party are weaponizing trauma triggers or appealing to character weaknesses.
"hop on the train, commit fully, or you'll be left behind" > fear of abandonment trigger
"pah, skill issue. my prompts on the other hand...i'm afraid i can't share them as this IP is making me millions of passive income as we speak (i know you won't probe further cause asking a person about their finances is impolite)" > imposter syndrome inducer par excellence, also FOMO -- thinking to yourself "how long can the gold rush last? this person is raking it in!! what am i doing? the miserable sod i am"
1. outlandish claims (Claude writes ALL the code) noone can seem to reproduce, and indeed everyone non-affiliated is having a very different experience
2. some of the darkest patterns you've seen in marketing are the key tenets of the gospel
3. it's probably a duck.
i've been 100% clear on the grift since October '25. Steve Eisman of the "Big Short" was just hopping onto the hype train back then. i thought...oh. how much analysis does this guru of analysts really make? now Steve sings of AI panic and blood in the streets.
these things really make you think, about what an economy even is. it sure doesn't seem to have a lot to do with supply and demand, products and services, and all those archaisms.
For all the technology we develop, we rarely invest in processes. Once in a blue moon some country decides to revamp its bureaucracy, when it should really be a continuous effort (in the private sector too).
OTOH, what happens continuously is that technology is used to automate bureaucracy and even allows it to grow some complexity.
See, this is an opportunity. Company provides AI tool, monitors for cases where AI output is being fed as AI input. In such cases, flag the entire process for elimination.
> The edges are where interesting stuff happens. The boring part can be made more efficient. I don’t need to type boring emails, people who can’t articulate well will be elevated.
I think that boring emails should not be written. What kind of boring emails do you NEED to be written, but not WANT to write? Those are exactly the kind of email that SHOULD NOT be passed through an LLM.
If you need to say yes/no. You don't want to take the whole email conversation and let LLM generate a story about why you said yes/no.
If you want to apply for a leave, just make it optimal "Hi <X>, I want to take leave from Y to Z. Thanks". You don't want to create 2 pages of justification for why you want to take this leave to see your family and friends.
In fact, for every LLM output, I want to see the input instead. What did they have in mind? If I have the input, I can ask LLM to generate 1 million outputs if I really want to read an elaboration. The input is what matters.
If I have the input, I can always generate an output. If I have the output, I don't know what was the input (i.e. the original intention).
when i pass my writings through ai the output is generally only marginally bigger than the input, and it derisks things a lot making my prose a nice beige.
This is a strange comment because, this is literally the world that we live in now? We just assume that everyone is vouched by someone (perhaps Github/Gitlab). Adding this layer of vouching will basically cull all of that very cheap and meaningless vouches. Now you have to work to earn the trust. And if you lose that trust, you actually lose something.
But testing in Python is so easy to abuse as LLM. It will create mocks upon mocks of classes and dynamically patch functions to get things going. Its hell to review.
reply