Also, for some reason Optional[T] became deprecated, just as the ecosystem finally embraced types ~3 years ago.
In fact, one my company's greenfield projects decided to use TypeScript instead of Python for the [surprisingly] more consistent tooling, and the fact that the big LLM providers all have official TypeScript SDKs anyway. Also, for agentic coding, LLMs don't seem noticeably worse at TypeScript than Python.
My experience can be summarized as:
- for some reason we need 2-3 static analysis tools just for typechecking
- no tool understands each other's comment directives
- each tool reports a different error in your codebase
- even big libraries (e.g. matplotlib) make half their functions return Any
- you'll be tempted to silence the "partially unknown type" warnings, and you'll have to do it for each tool that's running.
Deprecated in favour of `T | None` exactly because of that embrace. It's cleaner, more consistent (you can `T | U` arbitrarily), and helps slim down the `from typing` imports.
Optional[T] is now T | None. Means exactly the same thing but doesn't require an import. Support for the older syntax presumably won't be removed for a long while regardless.
I think Python is probably good for many things, including scripts and as a starter language, but I don't understand how anyone can stand writing large software systems in it.
> for some reason we need 2-3 static analysis tools just for typechecking
I don’t follow: you need one type checker, of which you have several options. It’s arguably not ideal to have more than one option, but you should never need to run more than one.
- no tool understands each other's comment directives
In general, all type checkers in Python support the `type: ignore` directive, since it’s standardized.
> each tool reports a different error in your codebase
This is a real problem, but I think you can avoid it (like most people do) by not mixing different tools that do the same thing together.
To my understanding, you’d have the same problem if you combined (e.g.) biome and eslint in a JavaScript codebase.
I have to use more than one Python type checker because there is not a single one that works. Not only different tools catch different issues. They also have different bugs, and different configuration requirements. Different teams have different preferences.
It's a nightmare. If Python taught me something about typing is that a language that doesn't have a clear definition of types in the reference implementation, it will never get it fixed with external tooling.
It's a complete ergonomic travesty that Python doesn't have one.
I personally prefer the fake typing in Python because it fits well with our defensive programming style with very low abstraction and little to no adherence to DRY. Since Python naturally force you to deal with runetime assertions rather than getting you to do compiletime checks that then don't actually offer any form of safety at runtime. Which is obviously not a very technical argument, but it just feels a lot cleaner rather than having to juggle the two.
I don't get how uv regularly gets recommended without any note about this.
Can you say more? uv should always tell you if a wheel build fails, unless the build backend (which uv doesn’t control, unless you use uv’s own backend) decides to silently ignore a wheel build. This would be a bug in any given build backend IMO.
This is an unfortunate complexity in Python packaging: something like `uv build` can dispatch a wheel build for you, but the actual code that gets run as part of that build is often third-party build backend code that uv itself has little to no control over.
The main difference is that in the JS ecosystem it is all installed at the project level, you don't need anything globablly installed besides the runtime and package manager (and even the package manager can be auto-installed as well if you set it up that way).
[1]: eslint, biome, prettier, scass linter, graphql-codegen, tsc, tanstack-router codegen. That I remember, might be more (although codegen might not be considered static analysis, it is needed for static analysis).
So same as JS then.
I have switched type checker recently in my own Python projects from zuban to ty. ty seems to work better, and is not a one man show/bus factor of 1, though I respect the work that has gone into zuban by its creator. But ty doesn't understand mypy configuration in pyproject.toml ...
I imagine switching a type checker in a bigger project and with more people involved to be a bit of a PITA, until everyone has adjusted their development environment/tooling. Best one can do is research beforehand which tool suits one best, test it, and then stick to it, unless it has unbearable failures.
Rather I prefer not to be in the same spot I was in 1999 - 2001, with Tcl, and every now and then rewriting code into C, for the application to actually deliver within the performance deadlines.
Python is the only mainstream dynamic language where runtime support for dynamic compilation is such an hassle, where the alternatives do exist, yet are mostly ignored.
I agree that short compilation times are very desirable, but I do not see why Python must be the solution for that.
I do not know whether Haskell can be compiled quickly, but from my experience, I am very certain that short compilation times are easily achievable for languages with good static type checking, especially with compilers that have different options that allow choosing between fast compilation and heavily optimized compilation.
An optimized compilation may require a much longer time than a fast compilation, but that has no relationship with the programming language used in the source text, but only with the intermediate representation used by the compiler and the target CPU ISA. Usually, if you compare the compilation times of multiple programming languages, all the compilation times with fast compilation options are much shorter than all the times with high-optimization options, so the programming language choice may be less important than the chosen compiler and its command-line options.
When you try to optimize a project by generating many variants with a LLM, I doubt that all those variants will be generated from scratch, completely independently, even if only for the reason that when using a commercial LLM the cost of a completely new variant will be much higher, by requiring many more tokens, so whenever possible it is preferable to generate other variants by just patching previous variants.
Whenever a variant is generated by editing a previous variant, incremental compilation can be used, which should be pretty much instant on modern computers.
There is also this
> How do we make library docs full of copy-pastable, realistic examples, not just beautiful types?
Which is useful for humans as well as agents.
Haskell indeed has a very bad track record of documenting its libraries. For many people, just having the function signatures is documentation enough.
Rust is equally bad at compile times (if not worse) but its standards for documentation is at another level
You make a fair point about a documentation gap. The first step is always defining the problem. Wanting to add lots of realistic examples sounds like a wish for more tutorial or beginner-friendly content. Do you see the problem differently?
On the other hand, a given pure function type only has only so many possible implementations — why tools such as djinn and MagicHaskeller exist or why Hoogle is actually useful, unlike the horror of searching for every `void (*)(const char *)` in C.
https://hackage.haskell.org/package/djinn
https://hackage.haskell.org/package/MagicHaskeller
From that perspective, Haskell docs tend to be more expert-friendly — perhaps a rationalization, granted — which seems ideally suited for an in-IDE model to help bridge between developer intent and typechecked code. However, this comes at the expense of putting in the reps to rewire the developer’s brain to think in functional terms and the resulting mind opening and horizon expansion to think new thoughts she wasn’t capable of even considering. In these days of LLMs, fretting over that particular opportunity cost may be thinking nostalgically about the loss of craftsmanship in fine, well-balanced buggy whips.
In the limit now, will all programming be strictly literate?
OCaml enters the room...
There's no reason to read the code anymore.
The LLMs produce it inhumanly fast.
They can usually debug and fix problems faster than Haskell can compile the project.
Meanwhile the open source communities are in basic denial about AI, so trying to change things by making compilation faster or advancing a Haskell interpreter are going to meet with fierce resistance.
Whatever. Give up. Just ship Python.
I sympathize.
have you tried?
i use elixir and carefully watch the agents and its very seldom making typing mistakes (elixir is in-between, it's typed but only as a checker).
ultimately typing doesn't help as much because it's nonlocal information. if the system can locally infer what the shape of functions is, it's way better.
- I do not watch agents, at all. Rust and Typescript. When I use Typescript only I have some guidelines so that we build the stack to be as strict and type driven as possible.
I don't personally use Haskell for anything, but I use Lean and occasionally some other languages with expressive type systems, and like you I've found it to be a pretty great experience for working with LLMs. But I've also experienced what the author is talking about, with languages that sit at different points on the type system spectrum, regarding a languages ecosystem/infra layer becoming a bottleneck. I don't think it's ultimately about the type system but the broader ergonomics of the language/ecosystem.
So I think his criticism is less than expressive type systems are a pre-LLM concept, and more that Haskell has an individually bad "agentic coding story".
so does this mean that the LLM writes code that is so good that the compiler does not find any more errors?
or is it due to the nature of haskell that makes it hard to write bad code to begin with?
or just that because the haskell compiler catches more errors there is less broken haskell code for the AI to train on?
and what does that mean for the switch to python? if the python compiler/interpreter doesn't catch as many errors do we even know that the code is good?
or is this more like the belief if the LLM can generate good haskell code, surely it can also generate good python?
what's the solution here? speeding up the haskell compiler? if that were easy, would it not already have happened?
personally i still don't trust LLM code generation. i didn't learn haskell yet, but what i hear about it makes me more likely to trust that LLMs can generate good haskell code than python.
i believe the future in LLM code generation is code that can be proven to be correct. proving code correct has been a research topic at some point.
It has been an area of active research for 40 years. But almost all the research returned the null result, meaning that the program proving didn't improve code quality (basically it didn't work). Yet somehow a group of programmers, usually fresh out of academia falls for program proving each generation. Strong types do really help but you need a good compiler which is sometimes lacking in the real work cough Scala cough. The problem with strong types and program proving is that the juice just isn't worth the squeeze meaning the extra time taken doesn't result in reduced debugging time or improved code quality. I don't think that changes with LLMs. It just exposes the flaws more quickly.
Zuckerberg became rich on php, which grinds my gears almost as much as the genocide thing.
In fairness, my peers at $corp used to ship while I was sad-Wojciech "no you can't create tech debt". The universe is cruel to people who care.
The problem has always been:
- It's extremely labour-intensive, and even small changes to the code can require an enormous amount of new proof work.
- The skills required to formally verify software are very different from the skills required to write it, and the set of people capable of doing so is much smaller.
- Code has to be written with verification in mind, against a specification that expresses invariants that can actually be verified, and that is coherent enough that you can derive useful high-level properties of the system from it.
You needed two teams — verification and development. Verification would always be behind the development team, while also having to feed requirements and design changes back to the developers. Everything slowed to a crawl.
AI changes this. A coherent, verifiable, useful specification isn't easy to write — but it's far easier to write than the software itself, and AI can do most of that work: both drafting the spec and proving that it's consistent and that the high-level properties you actually care about follow from it.
More importantly, a high-level spec is far easier to read and reason about than the reams of code required to actually implement something. Which means:
- AI does the grunt work of writing and proving the spec; humans only have to carefully review that high-level artifact.
- AI writes code it must also prove conforms to the spec, so humans can be assured it's correct without babysitting the AI.
- Changes are driven top-down: evolve the spec first, then have the AI fix the implementation and re-prove conformance.
Our (very, very large) company is rapidly going all-in on formal verification across projects we never would have dreamed of verifying before; the velocity hit and the man-hour cost were only worth paying for truly critical infrastructure.
I suspect you’ve nailed the answer: it’s probably not easy, although it’s also possible that it just hasn’t ever had a lot of attention paid to it because it’s been generally fast enough for their user base?
But in commercial production environments, CI pipelines tend to want to build everything from scratch every time, and that slows everything down. Rust has the same issue. Both languages, by default, compile all their dependencies from source, rather than obtaining precompiled artifacts from a repo the way some languages (like Java) do. And their compilers are slower than e.g. Go's. As the article mentions, various kinds of caching can help with that, but that's extra stuff you have to manage and deal with.
I'm not sure this is a bad thing, though. Haskell co-creator Simon Peyton-Jones coined the unofficial Haskell motto, "avoid success at all costs". I tend to agree with that. It would be difficult for Haskell to maintain its conceptual edge if it were a mainstream commercial language.
Refactoring spaghetti has become easier in the LLM era because it can just read all the code, and there is now a skill floor on the programmers that kicks in somewhere relatively high. The benefits of type systems might have suffered because of that.
They're absolutely huge for this, but you have to write code to take advantage of the guarantees that the type system can offer.
As Yaron Minsky at Jane Street put it, "make illegal states unrepresentable". Stronger type systems make it possible to make more states unrepresentable. You end up with what amounts to static debugging - you debug your code at compile time.
Sure, it's still possible for runtime bugs to occur, but entire classes of bugs are eliminated, plus it becomes possible to have static assurances about program states about things that most language don't even try to express in the type system, like security.
In Python, a "class" is simply a dict under the covers and (by default at least) you can add attributes to it after definition (as well as things like properties). So it's difficult to reason about what the fields are at any given time. And that's assuming people USE classes! I've seen code where all the state is in one giant ever-changing dictionary and you have to pull out a debugger just to figure out what's IN the thing! God help you if you mispell a key!
Maybe you work with better quality code than I do, but I find Go's type system a lot easier to reason about than Python's.
I would quibble though that Python's is actually pretty good at this point, and, despite what the below poster is saying, straight-forward to set up and use. I am still perplexed that the author chose Python over Rust or Scala or TypeScript though, especially given they presumably want to migrate a Haskell codebase.
I'd be interested in hearing/discussing more about this. I was very surprised, when I embarked on my side project, that Rust's options for SQL ORMs all seem so weird.
I think what you are referring to is the derivation of FromRow and stuff with SQLx, right?
This means you have rock solid assurance your program won't have database syntax errors at runtime. Basically it's just one less thing to worry about, and it's ideally suited for LLM-generated code.
I remember that being one of its "pros" when I was going over options with the LLM, but now, I just have a bunch of raw SQL strings in my codebase and sqlX's main use in that project, off the top of my head, is just instantiating Rust objects from the raw results with `FromRow` (it's probably doing more than I realize; I am not as connected with the code as I want to be, using LLMs to move fast to launch a couple features before revisiting a lot of the mess).
Most bugs aren't type issues until you make them be type issues by expressing some business invariant in types.
Refactoring makes an exception not being caught the same way as before ? Type issue. Mixing up some ids ? Type issue. Etc.
Now that can also be emulated with extensive tests. But isn't that a concern for OP as well ?
perl another offender… is it a hash? is it an arrayref? over time you get it right, but by trial and error and looping. json suffers this too, arrays different from strings, different from numbers etc, but opaque until checked and liable to change
Which is basically what people liking dynamic languages have said all the time - types is only good as long as the overhead they bring doesn't cost more.
Also, you can just compile less frequently.
But hey, if LLMs are what drove this person from Haskell to Lisp then all the power to them!
I didn't see Lisp mentioned in the article. They moved to Python. Which is certainly a choice.
Besides, “machine code generation” is not a fundamental part of Lisp. Some of the most famous implementations were pure interpreters.
Toy lisp interpreters are easy and fun and everywhere, often written in lisp itself, but the same books that teach them go on to teach compilers. SICP chapter 4 is the metacircular evaluator, chapter 5 is the compiler. That ordering is the point: interpretation is the pedagogy, compilation is the destination. Lisp was designed to be easy and efficient to compile (so was Self, and that tech lineage runs straight into HotSpot and V8). The archaeology cuts the other way too: the only famous pure interpreter was LISP I on the 704, and within two years Hart and Levin's Lisp 1.5 compiler (1962) became the first self-hosting compiler in any language, written in Lisp, compiling itself. Lisp practically invented compilation as we know it. And today SBCL doesn't even interpret by default: a form typed at the REPL is native code before it runs. Clojure compiles every form to JVM bytecode.
So antonvs, I'd genuinely like to know which famous implementations you mean, because the honest counterexamples aren't the famous Lisps at all. They're the embeddable extension languages, and I can speak to those from direct experience, not just archaeology.
A digression about Elk, since hardly anybody remembers it and Oliver Laumann deserves the recognition. Elk, the Extension Language Kit, was Laumann's Scheme out of TU Berlin, started in 1987 to be the extension language of the ISOTEXT document editor, released to the world in 1989. Its stated goal was "to end the recent proliferation of mutually incompatible Lisp-like extension languages": instead of inventing yet another one, you linked the Scheme interpreter into your C or C++ application and taught it your app's types and primitives. Emacs-style extensibility as a kit, years before Guile existed. And it was a pure interpreter on purpose. Laumann said it plainly in the USENIX paper: performance was "uninspiring (no compiler is available)" and it didn't matter, because any bottleneck got recoded as a C primitive. The compiled half of an Elk system was the application itself. It shipped that way in real products, including the TELES.VISION video conferencing system, which used Elk continuations as its multithreading mechanism. Elk's embedding design went on to inspire other interpreter authors, including Matz for Ruby.
I hacked on Elk back then and wrote the SPARC port, credited in the Elk 1.2 release notes, and it taught me why "just an interpreter" is never just an interpreter. Elk implemented full call/cc the brute force way: capture the registers setjmp-style, then copy the entire live C runtime stack aside onto the heap, Scheme frames, Xt frames, qsort frames and all, and copy it back whenever the continuation is applied. That works on machines where the stack actually lives in memory. On SPARC it does not: the register windows keep the most recent frames cached in the register file, invisible to any memcpy of the stack. So before copying you trap into the kernel to force every window to spill to RAM (ta ST_FLUSH_WINDOWS), grab the stack while it's momentarily telling the truth, and pray it all lines up with the calling conventions when you copy it back so returns don't sail off into the weeds. Frightening! As hairy as anything inside a code generator. The "simple interpreter" strategy relocates the machine-level work, it doesn't remove it: stack layout, register windows, calling conventions, incremental linking against the running executable so compiled C extensions could be loaded into the interpreter, unexec-style dumping of customized interpreters back out as executables. Compiler and linker work in everything but name.
And look what happened to every extension language that survived: Tcl grew a bytecode compiler in 8.0. Emacs Lisp went from bytecode to native code. Guile, after fifteen years as an interpreter, got a compiler in 2.0 and a JIT in 3.0. Interpreters are how these languages are born. Compilers are how they grow up. Even Ousterhout's own student proved it during the Tcl wars: Adam Sah's Berkeley thesis (TC, 1994, first reader John Ousterhout) got 5-10x by caching parsed representations of values, essentially the design Tcl 8.0 later adopted, and his Rush language with Jon Blow (yes, that Jon Blow) compiled a Tcl-feel language to Scheme at 50-300x stock Tcl. RMS cited Rush by name in the GNU extension language announcement. The Rush source is lost now, which is a small tragedy.
Since Guile came up: Guile existed before the Tcl war, not because of it. Tom Lord had already forked Aubrey Jaffer's SCM into an embeddable library at Cygnus in 1993, called it GEL, and talked RMS into blessing it as the official GNU extension language. The rename to Guile came after trademark trouble, partly because it sounds a bit like "Guy L." The war started when Sun swaggeringly declared Tcl would be "the ubiquitous scripting language of the Internet" and a colleague skunked Tom's Scheme-based GDB GUI project with a quick Tcl/Tk one. Tom vented to RMS, and shortly afterward "Why you should not use Tcl" landed on comp.lang.tcl. Who actually wrote and posted it stayed deliberately murky: the Cygnus NNTP server showed it coming from Tom's dormant gnu.org account, some archived messages have RMS in the From line and Tom's signature at the bottom, and Tom would only ever call it a semi-prank by "the Scheme underground."
Why didn't Guile take over the world? Partly because the plan was always grander than the resources: the idea was one Scheme engine executing many surface languages, including a repaired Tcl and eventually Emacs Lisp itself. But Emacs Lisp turned out to be immovable. It isn't just a Lisp, it's a deeply weird Lisp: dynamic scoping by default, buffer-local variables, text properties, semantics tangled into the C core of the editor. There were decades of plans and experiments to put Emacs on Guile, and Guile 2.x even shipped an Elisp compiler on its VM, but the ecosystem never budged, and in the end Emacs went the other way and grew its own native compiler instead. Guile's own arc is instructive: Andy Wingo gave it a real compiler in 2.0, a register VM in 2.2, a native JIT in 3.0, and it found its killer app in Guix. Not the ubiquitous extension language of GNU, but a living vindication of the design argument Tom was making in 1993: a general Scheme engine underneath, other languages and applications on top.
Tom was a good friend of mine. We argued for decades, about Tcl and Tk and everything else, the way you can only argue with someone who has actually built the thing being argued about. He died suddenly in 2022, and the field is short one fiercely original mind. His body of work and his philosophical posts about extension languages, universal engines, and who gets to control the substrate are a rich source of wisdom that hardly anybody mines. His account of the war is him on the page: brilliant, funny, hard on himself, and fair even to the people who did him wrong. Read it, and the detail I love most for this thread: even Tom, patron saint of the embeddable Scheme interpreter, compiled the performance-critical parts of his Scheme editor with the Hobbit compiler. Nobody stays pure.
An Account of the Tcl War, by Thomas Lord:
https://web.archive.org/web/20110102015130/http://basiscraft...
Glenn Vanderburg's archive of the original flamewar:
http://vanderburg.org/old_pages/Tcl/war/
Elk: The Extension Language Kit, Laumann and Bormann, USENIX Computing Systems, 1994:
https://www.usenix.org/legacy/publications/compsystems/1994/...
I was on all sides of that war at once. Tom and I had heated arguments because I was shipping on the other side: I ported SimCity to Unix with Tcl/Tk, and Tk was the driving factor, not Tcl (and Tcl's weaknesses were real, and Tom and RMS's arguments against it were largely correct: I used Tcl/Tk extensively and pushed them in ways they were not originally designed to be used; multiplayer Tk was a challenge, I had to fix multi-head bugs in X resource management and Tcl UI menu and other mouse-tracking code, and I built a fully functional multiplayer coordinating voting collaborative game on top of it instead of just getting it to draw googly eyeballs on two screens at once; I am painfully aware their arguments are right).
But having an extension language at all is a higher-order bit than having a perfect one, especially when you are designing a GUI toolkit, whose concerns partially overlap those of an extension language. Look at what the X toolkit intrinsics and the toolkits and window managers built on top of them abused X resources for: stringly typed nano-domain-specific languages and semantics and protocols between differently named resources in the same file. Key and mouse bindings, user configuration, fonts and colors and measurements and dimensions and scales, state machines, commands, preconditions, postconditions, parameter bindings, and and and I could go on forever. All of that is what a real extension language is for. Tk worked so well precisely because a real scripting language (no matter how arguably weak) was part of its design from day one, so it never had to badly reinvent one, which is Greenspun's Tenth Rule, the curse of every extensible app before and since.
And the ecosystem kept proving the point: Python's tkinter still embeds a Tcl interpreter with each Tk root, assembles Tcl/Tk command strings, and passes them through _tkinter to be evaluated, and you can register Python functions as Tcl commands and trampoline back across the boundary (the same class of plumbing obsession as SWIG: who owns the glue, and what shape is the trampoline?). Tk never really grew a clean pluggable "bring your own scripting language" extension point; Perl Tk, Ruby's tk.rb, and the rest generally still talk Tcl to the toolkit. The design win is simpler and more Self-like than that: assume a scripting language, any scripting language, is available at runtime, and the toolkit can stay small and SIMPLE.
Which loops back to one of the greatest pieces of language design literature the Tcl War left behind, one that historians and archaeologists will study for centuries: what even counts as a scripting language? Ousterhout's answer was the system-programming-language versus glue-language split, two languages for a large system, typelessness and strings as the uniform wire format between components.
Scripting: Higher-Level Programming for the 21st Century, John Ousterhout, IEEE Computer, 1998:
https://web.stanford.edu/~ouster/cgi-bin/papers/scripting.pd...
I even demoed multiplayer SimCity to Ousterhout in his office at Berkeley.
And then Sun, having hired Ousterhout and his team, then anointed Tcl the ubiquitous scripting language of the internet, pivoted on a dime to Java, got its lunch eaten by JavaScript, a language whose deepest debt is to Self, which Sun also starved. The Self people who left applied the ideas at Animorphic, got bought back, and built HotSpot. As an OG NeWS developer I know exactly how it feels to be promised the world by Sun and left on the back porch in the rain, so I have nothing but empathy for the Tcl/Tk and Self teams. I've been posting these receipts for a decade:
https://news.ycombinator.com/item?id=28669698
https://news.ycombinator.com/item?id=25888450
And the one I treasure, a conversation I had with Tom in 2006, two years before V8, about running into Dave Ungar and the irony that JavaScript credits Self for its prototypes while missing everything else the paper was actually about. It's right there in the title: "Self: The Power of Simplicity." Simplicity first.
And behind that, the two hard tricks that made the simplicity affordable: compiling a radically dynamic language to fast native code, and keeping it debuggable at the same time. That last one is the sleeper. Anyone can pick two of simplicity, performance, and debuggability. Self's dynamic deoptimization let you debug fully optimized code as if it were running unoptimized, source-level, mid-flight, and that machinery went straight into HotSpot too. JavaScript took the prototypes, which were the cheap part:
Self: The Power of Simplicity, Ungar and Smith, OOPSLA 1987:
https://bibliography.selflanguage.org/_static/self-power.pdf
Debugging Optimized Code with Dynamic Deoptimization, Hölzle, Chambers, and Ungar, PLDI 1992:
https://bibliography.selflanguage.org/_static/dynamic-deopti...
https://news.ycombinator.com/item?id=33527618
Which brings this all the way back to Haskell and the article. timcobb joked about being driven "from Haskell to Lisp," but that move would arguably have been more defensible than Python, because the real lesson of seventy years of Lisp is that interpreted versus compiled was never the actual choice. Batch versus incremental is the choice.
andersmurphy and jwr have it exactly right about the REPL. Common Lisp and Smalltalk systems compile one function at a time, to native code, inside the running program, in milliseconds. Avi's headline brag, fixing a bug before the customer hangs up, was routine practice in Lisp shops in the 1980s. NASA once did it to a Lisp system running on a spacecraft a hundred million miles away: the Remote Agent on Deep Space 1 deadlocked in flight, and they diagnosed and fixed it through a REPL running on the spacecraft itself.
Lisping at JPL, Ron Garret's firsthand account of debugging Deep Space 1 in flight:
https://flownet.com/gat/jpl-lisp.html
giraffe_lady named the two axes precisely: speed and accuracy of feedback. GHC gives you accuracy at the price of batch-world speed. Python gives you speed at the price of accuracy. The Lisp lineage, from Hart and Levin through SBCL to the Self-descended JITs your JavaScript runs on today, has been demonstrating for decades that you can have both, if you design the language and the compiler to work incrementally, together, as one live system. And I can't resist pointing out that SBCL's native code compiler is named Python, and carried that name years before Guido picked it for his language. Lisp hackers have been getting sub-second feedback from Python since the 1980s. Theirs emits machine code.
CMU Common Lisp, whose compiler has been named Python since the mid 1980s:
https://en.wikipedia.org/wiki/CMU_Common_Lisp
Oliver Laumann and Tom Lord were exploring exactly these tradeoffs with practical, load-bearing, shipping code while most of the industry wasn't looking. The agents everyone in this thread is arguing about are just the newest programmers to want what Lisp hackers always wanted: a fast, honest conversation with a running system. It would be a shame if we rediscovered everything except the part where it compiles.
I didn't get a chance to post about my memories of Tom in the hn discussion of his death, because I hadn't really processed it in time to comment, so I'm writing them up now and will publish more later.
But for now, a few memories of Tom Lord, as promised.
In April 2010 Tom posted me a link out of nowhere: "That reminds me (why? I dunno) that - hey: hate man made the news."
Homeless ex-reporter opted for Berkeley streets (SF Chronicle, April 2010, via archive.org):
https://web.archive.org/web/20100413124819/http://www.sfgate...
Hate Man was Mark Hawthorne, a New York Times reporter from 1961 to 1970 who opted out of normal society and spent decades on the streets around Telegraph Avenue and People's Park. He wouldn't talk to you until you told him you hated him. From the Chronicle interview: "It's a new way of hating. It's about being straight with people... My idea is to be straight about negative feelings that we all have, which is what hate is, and then you can have a real conversation."
I replied that I loved Hate Man (but that kind of sounds weird, like I totally missed the point), and that I was honored he'd hated on me once and wished we'd discussed it more deeply, and that he seemed like one of the most sensible people walking around Berkeley.
Then Tom wrote something I've been thinking about ever since. He said that back in the day he'd had some really, really bad days, real existential crises, and that "Hate man and several of the brothers in the park basically saved my life." Hate Man by getting him to play his "let's lean on each other game," the brothers in the park by being good listeners to mumbled words, often enough good counsel, and by whooping his ass at street chess. He described how, as things shut down for the night, Hate Man would gather the rough itinerant runaway kids of Telegraph up in a circle in Sproul Plaza, and they'd all chill and keep safe and share warmth and make peace, "all in the name of Hate. Telegraph was very much a zen temple. Hate man a famous monk."
He also corrected the Chronicle's timeline. The article placed the start of Hate Man's street life in '86, and Tom called bull: Hate Man was the first resident of Berkeley he ever heard of, in 1983, in Massachusetts, from someone who was basically trying to encourage him to run as fast as he could away from the preppy high school that G. W. Bush went to: "There's this guy there. And he's really cool. He likes to shout 'I hate you' at people. And he won't talk to you unless you first tell him that you hate him, too." It made no sense to him at the time. Years later, he said, arriving in Berkeley was like walking into the landscape of a favorite children's book he'd read years before. Many of the familiar characters were there.
I told him that was beautiful.
He replied: "I hate you."
I replied: "Go stick your head in a pig, beeeyatch!"
That's a lyric from Share and Enjoy, the Sirius Cybernetics Corporation jingle from The Hitchhiker's Guide to the Galaxy:
https://www.youtube.com/watch?v=_wSBC5Dyds8
That was sixteen years ago. Tom died suddenly in 2022, and last week I finally took my turn again, on his memorial Facebook page: I HATE YOU!!!!
Hate Man would approve. Be straight about the negative feelings first, and then you can have a real conversation. I hate you, Tom. I miss you.
And if you want one minute of the town where all of this made perfect sense, Tom Lord's Berkeley, here is a shirtless dude riding a cow in the wrong direction up Telegraph Avenue:
Cows In Berkeley? Moooo:
https://www.youtube.com/watch?v=10JSPdTDFsI
I love that town. It suited him perfectly.
If you push through that you end up with code written in a language people have used for formal proof (seL4 model is Haskell) and deployment wise a binary that +/- libraries you depend on ought to be reasonably portable.
I'm very surprised anyone would want to go the other way. Same ecosystem pain, plus you need to start shipping interpreters or containers, plus the language just doesn't really compare.
It could just be a coincidence of statistics, but it does cover every Haskell user I know (needless to say, these people are much smarter than I am).
Start the countdown timer for how long it takes them to discover that was a mistake.
Nothing to do with Haskell, but good grief, LLMs do not in any way, shape or form save you from the deep, unfixable problems with Python.
At the very least you need all the static checking machinery like Ruff, Pyright, and hefty unit tests that take the place of typechecking if you don't want obvious failures to only show up in production.
I had this recently with an ML training pipeline, where Python is essentially forced on us. A dynamic error occurred after 17 hours of training - something that a real type system could have easily caught.
The solution that the LLM came up to prevent this in future was a complicated Enum-based system that just made me wish I could use a real programming language.
I have the impression that Python basically wins by default in those spaces due to the lack of many good libraries in other languages (except for, like, C++).
But curious if this is just a very outdated view of the world
The type safety we gave up hasn’t been noticeable in any concrete way yet, especially considering our test coverage has never been better.
But I do think what benefits LLMs is the speed and accuracy of feedback. Type systems cover the accuracy part, but haskell was killing them on speed. It seems like a strange choice to go so far the other way on accuracy when there's a lot of languages in between. But I'm not familiar with the project so not in a position to call it.
It's not also really about expressiveness IMO. I've found LLMs to be best with more constrained type systems: they are better at ocaml than they are at typescript.
But I'd look at people a bit oddly if they said: 'We didn't want to set up CI caching and compiled languages took 30 minutes per run so we changed our entire codebase to python'.
Maybe it makes sense for them, and caching across dynamically spawned VM's is admittedly a harder problem which most build systems aren't great at, but still. I can easily believe that getting build caching to be reliable would be a lot of work, but is it more work than a full rewrite of a significant codebase?
Or if staying on Linux, JVM snapshots.
Secondly, there are several ways how Java source code becomes machine code, depending on which JVM and JDK is being used, not taking into account the ART cousin.
When the potential set of behaviors you could write a program to have is infinite, but the actual behavior you want is singular, a programming language is more importantly defined by which ones it eliminates up front than which ones it lets you write (assuming it lets you write the one you want at all, but that's almost always going to be the case for most general purpose languages). Bugs are just false positives in this framing, where the program you wrote seems like the one you wanted, but there's some divergence between what you thought you were getting and what you actually got, and catching some of those up front is a huge part of why type systems are so useful.
Because a lot of times it's missing the way of making the needed steps for the conversion (or it's just not obvious)
Sometimes it needs some nudging
Also this comment is a bit generic, it can also apply to cases where it's not an obvious "type check" but a redundancy that needs to exist but the AI can't get around
One approach would be to not use LLMs.
& ultimately this is why Java/JVM keeps winning & to a lesser extent Typescript.
you benefit from the robustness of the JVM, compilation is fast, the language is fast enough.
to apply to Javascript/Typescript - good enough, fast enough though will not reach performance of JVM.
I am not following. Why does it apply to Ruby? Ruby has likely 10 - 100x the size of open sources project and momentum behind it.
I suspect that the cost of long compilation times (preLLM even) was actually quite high and the author is discounting/excluding those other factors and focusing on LLM feedback loops primarily as their justification.
As an aside short feedback loops are important, but the article ignores one major reason why: humans learn most effectively when the time between action and feedback is reduced because the contents of our working memory degrades over time (take the explanation with a grain of salt). LLM has no such restriction, so the only thing that matters is that the codegen can keep up with the demands from the actual product.
For some time now it’s felt clear (or at least extremely) compelling that agents need fast compile times in order to be effective, especially when you’re working in parallel. But the other thing that has felt just as obvious is that agents need strong type systems and narrow guardrails in order to constrain their outputs. These two things felt clear enough to me that, like the author, I wanted to choose a language ecosystem that maximized them. There _are_ languages that both have expressive type systems _and_ fast compile times. I wonder if the author investigated any of them, before deciding that no compilation time at all was acceptable.
In my case I landed in OCaml. I think there are other options in the space—Go if you want less typing but faster compiles; Rust if you want more types but slower compiles. My mostly vibes-based evaluation landed on OCaml, and I’ve been pretty happy with the results.
- None of us are experts in Rust, and we're all solid at Python.
- Rust felt like an under-correction for what we wanted (get all friction in front of the LLM out of the way).
- Our high-performance stuff is not being migrated at this time (Scarf Gateway), so we're just talking about basic CRUD backends here. Basically any language will work.
Whereas Python just interprets and gets off to the races.
Feels like we had this discussion years ago as humans..the false promise of dynamic languages.
it's a shame scarf is struggling so much they are pinching pennies :/
you don't need to be, you can learn Rust or whatever way-better-than Python language as you use it with an LLM! it's an amazing process.
On the contrary, for a small rust project, I had to clean out 180gb of cargo nonsense from the last ~3 days worth of compiles on a single, narrowly focused topic branch.
The library situation might be funky, but I'm also learning Lean 4 by hand. The tooling & lsp integration is lovely.
I also went through quite a process to select a language to work with LLMs[1] before settling on OCaml.
I am not unhappy with the choice and find it works quite well, with relevant skills loaded, but I am always interested in others’ experience and understanding what they’ve discovered works well.
I read the second paragraph of linked article as saying close to the opposite of that, particularly,
"the model can often avoid the mistake before the compiler ever sees the code. And as the models get better, the relative value of catching every possible issue at compile time changes."
In other words, LLMs are much less likely than humans to make dumb, fat-finger mistakes, and, when they do, are able to catch and fix them more quickly, ergo the value of type checking has fallen.
Everything in the prior sentence is, obviously, highly debatable. But it felt like part of the premise.
All of these I run in a cli that has automatic LSP in it so that's a huge factor too. The agent is automatically told when there would be compile time errors as well as linter issues.
I would also add Kotlin, Clojure and F#.
Scala not really as the compilation is not much better, and since the Scala 3 reboot, the ecosystem doesn't seem to be doing that well.
The market opportunity for Haskell on the JVM is gone, although they are doing cool stuff with capabilities.
IMHO there are a bunch of reasons
* Scala 3 was a Python 3-esque disruptive event
* Perception as an overly-complex kitchen-sink inclusive language
* Kotlin took up the "better Java" mantle
* Rust became attractive to devs not committed to the JVM
* Go became less sucky
* SBT is disliked and Metals was buggy and unstable (and still slow)
* Suffers from the "we won't be able to find devs" problem where Kotlin, Rust and Go don't
* A bunch of community drama drove some devs away to Rust
I say this as someone using Scala 3 with ZIO effects system and Mill build tool and thinks it is fantastic.
There were other breaking changes as well.
https://docs.scala-lang.org/scala3/guides/migration/compatib...
This naturally broke all the tooling.
Then you have Metals for VSCode InteliJ plugins, while the Eclipse plugin was dropped.
InteliJ plugin is much further than Metals, however there is the conflict of interests with pushing Kotlin instead.
Meanwhile most Scala shops have pivoted to also give feature parity on modern Java, and Kotlin, thus reducing the interest in using Scala in first place.
However as mentioned, they are doing cool stuff with capabilities at EPFL for Scala 3.
https://virtuslab.com/blog/scala/introduction-to-scala-3-che...
PS The Scala team should have been fixing their type inference engine which lacked some important features and needed a better (Horn clause based) architecture. They did a complete rewrite instead.
*https://signalsandthreads.com/performance-engineering-on-har...
At least from the sidelines, it looks like they've built/forked most parts of the OCaml ecosystem to make it work (OxCaml, Dune was built by then, custom profilers, custom networking hardware, etc).
Of course a lot of their work is open source and can be used, but I wouldn't assume that it would be worth the effort to do something similar with OCaml yourself. Maybe it is! LLMs go brrr and all that.
The amount of certainty random people have that LLMs have already revolutionized software development seems to be directly proportional to the media awareness of the AI companies finance unsustainability.
And that is not even considering how often the agent needs to run tests to get it right.
Regarding the first, I think you're probably right, but then again, if there is a 15-minute base cost, it's hard to amortize that through fewer incremental runs of the compiler.
(Which isn't to say that I think they are doing the right thing.)
Ironically, this is also what makes them shine with LLMs, the LLM has access to the running program and can modify it while it's running to get feedback instantly.
Complex type systems are cool. But, they are not free. I say this as someone who's first programming language was Haskell.
Amazing. How did that happen? Is it true that functional programming is only counterintuitive because almost everyone starts out with an imperative language?
How much faster? IOW, what's the difference (in minutes/seconds, not in percentages) between vibing Haskell and vibing Python?
Slow compile times should have been a deal breaker for how they impacted human coders. LLM coding just makes the problem more stark.
I have yet to experience a RHEL major version change that did not blow up all my tiny simplistic python scripts. I see the following options for using python: * run inside the container it was developed in * build your own python interpreter and environment and libraries and never use python pieces from the OS (i.e., act like a container without using one) * keep different versions of the code for different OS versions and use AI to rewrite all the code for the new OS version
Start to consider third party dependencies, and none of those feels tractable without an AI assist.
I've dabbled in C device driver code and kernel version differences were my only problem, not C. My perl scripts never break. My bash scripts rarely break. My dabbling in erlang didn't suffer from language version differences. My little elisp hasn't broken. Only python has inflicted this level of version pain. I have a colleague who says java has the same version pain as python, and from what I've seen from Jenkins maintenance he may be right, but I don't have colleagues who want to read java code, so I haven't written my own.
We're a Haskell shop (and have been for over 10 years now) and are finding agentic development with Haskell to work pretty damn well.
Cold compile times in Haskell are painful indeed. Our development practices don't really cause us to do that much - even with agents.
It's unclear to me if the development practices at Scarf that cause them to hit this pain often are worth it if it means giving up Haskell because the compile times are too bad. Maybe they are, but I don't think so.
For this reason alone I want to like Bazel. But at the same time it has like half a dozen caches for different purposes and doesn't feel generally elegant. It saddens me that (a) cargo can't do this afaik and (b) its hard to package Bazel packages under nix. I'm not sure what other system has a shared unlocked cache.
Since Opus 4.6, LLMs have been pretty clever at using fancy types with libraries like Servant and Beam. The expressiveness of the types, combined with feedback from the compiler, means that agents converge quickly to something that works. I don't think I've noticed agents having to run the compiler so often that compilation speed is an issue.
Haskell compiles slowly, but we have not found that to significantly inhibit AI-supported dev so far. We use ghci with `-fobject-code`, and our largest leaf module (10k lines of webserver request handlers) takes 8 seconds to `:reload`. To run stuff in it, the agent pipes `:reload`, or other invocations, into `ghci`.
Working on parts parts that are early in the module chain, such as our Prelude, makes the whole-project `:reload` take 50 seconds. Much of that could be avoided if we didn't suffer from the TH recompilation problem (https://gist.github.com/nh2/14e653bcbdc7f40042da3755539e554a). Originally I made a small GHC patch to hack that out (what this conservative recompilation protects from cannot happen in our project), which made the reload much faster, but the logic was changed in recent GHC so that doesn't work anymore.
In C++, we have some individual files (and thus compilation units) that take 45 seconds to compile.
Python is of course the fastest to (build+run). However, Python also has some problems with repeated runs: Once you have many imports, just starting the program (e.g. `--help`) takes ~2 seconds resolving imports. Do `import pytorch`, add another 2 seconds. For repeated runs, this can be a pain; Haskell and C++ are much faster for that, so they win when you don't have to change the code but repeatedly use already-compiled tools in a different way.
In all 3 languages, getting things to typecheck is very fast, because the agent directly reads the vscode LSP typechecking errors which are < 1s feedback. Claude Opus 4.6-4.8 understand all 3 languages very well.
I agree that GHC devs should focus most of their efforts on compile speed, and real-world pain solving. Some parts of that are indeed being done (e.g. newest GHC can write bytecode to disk to make the above workflow much faster, and codegen is by far the slowest part of compilation). But I think the focus should be even more on that. I consider most important and unsolved:
* solving Generics being slow to compile, especially for types with many constructors
* solving deriving classes being slow to compile
* solving TemplateHaskell causing too much recompilation
* doing staged compilation,
so that the next module can typecheck as soon as its imports are typechecked,
as opposed to waiting that codegen is done;
this unlocks a large amount of parallel work availability
Other real-world production things are being solved quite nicely currently:
* The new Haskell debugger
* Much better stacktraces
* Much better runtime introspection to debug runtime hangs etc
But I find it would not be a great language for the other parts. Its concurrency story is bad, its interpreter is very slow and immediately disqualifies Python when you need to do e.g. a line of code for every 1000 Bytes (say for streaming small data chunks), typing is very useful but bolted-on and not always correct.
Things where Haskell remains best-in-class is anything nontrivial-webservers, I/O, program correctness, the main "general purpose" programming, refactoring and long-term maintainability.
LLMs removed some of the difficulty of Haskell, e.g. some interesting ideas to make things safer with sophisticated types or mechanisms like TemplateHaskell are now MUCH easier to implement in a few minutes. Also Haskell's general stance of breaking backwards compat in fundamental places if it's needed to make the language better, thus causing moderate amounds of upgrade grind, completely evaporated as a drawback because LLMs are extremely good at fixing type errors to bring code up to date with the latest changes.
The LLMs are very good at explaining Haskell compile errors.
Learning Haskell and maintaining projects in it is now easier than ever.
This is doing a lot of heavy lifting here. How do you even know a mistake has been "avoided", and safely, for that matter? The kinds of mistakes the AI will miss are the worst kind: subtle logical mistakes, buried inconspicuously within boilerplate code, hard to detect or reason about, until one day you wake up to find that your database has been replaced by a rickroll video.
Also, it's odd that they moved from Haskell to completely untyped python ... python may be no Haskell when it comes to type safety, but typing+mypy/mypyc goes a very long way.
If you go from taking 2 days to write some code and 20 minutes to type check (which does seem long, don't get me wrong, but still) to 10 minutes to prompt some code and 20 minutes to type check, that percentage increase to me isn't enough to justify switching.
You're still almost 2 days ahead, and converting those 20 minutes to 20 seconds are not going to make you ship a feature appreciably faster. But those types stand strong and I don't believe they can yet be replaced by an LLM believing they're correct.
Having said that, I also think that Haskell should massively speed things up. Having strong types if nothing else should surely produce some amazing type-checking speed wins.
I would like to add one additional observation, since we have been using both Haskell and Python in production for a long time:
Haskell excels at platform work, while Python excels at product work.
Our infrastructure teams work in Haskell (and also Rust nowadays), while our product teams work in Python. This gives us the best of both worlds (in my opinion): fast and rock-solid infrastructure on the platform side, and fast development speed and quick iteration cycles on the product side.
This setup has worked well for years for us, but it remains to be seen how and if this is going to change as well in the new AI era.
My intuition is that type-safe languages with fast compilers are the best option. Maybe Go? I personally prefer Java just due to my experience running it in production, but am not sure there's many arguments for it over Go in a greenfield application. The other candidate would be Rust, but I worry about token efficiency and tool performance, I suspect it's not worth it for the runtime improvements.
All that being said, in this article switching to Python seems like a wild choice. Relatively poor performance, no compile time checking at all. Python's big selling point was developer ergonomics, which seems largely irrelevant now.
These are all just thoughts at the moment, I should try to find some evidence one way or another.
I reckon language choice matters more at the edges of economic activity where a specific language feature really does make the difference in the end product, but most activity that is leveraging LLMs now is more generic enterprise SaaS software.
Go is the perfect language for this new world. Its development loop is fast, it has types but not too much so, it is memory safe, it is easy to deploy and it runs efficiently enough for most line of business use.
Go was originally developed specifically for a world of interchangeable hard-working juniors, which is exactly what LLMs can scale up with only your budget as the constraint.
I'm wondering what advantage a hypothetical faster-compiling-Haskell would have in that world.
That is: Have a Haskell base system. Have a Python "development" version on which you iterate at lightning speed. But also, in the background, moving at whatever pace it takes, have an agent running that imports all the Python development changes into the Haskell version. Have nightly builds of the Haskell version to reap its benefits (issues caught by the type system, more efficient native code). They must have continuous or nightly processes to fix bugs in the Python code anyway, there is no way that all the things they ship "while still on the call with the customer" are always tested on the full test suite and always 100% correct.
And it doesn't have to be Python/Haskell of course. The "development" version could be a (hypothetical?) interpreted Haskell. I have no idea if ghci would be useful for this. Neither do I know if the 15-minute Haskell build time is spent in the frontend (so an interpreter would have to pay that cost too) or in code generation or linking (which the interpreter wouldn't need to care about). Anyway, these are things I would think about before I did what the OP did.
Now I am using Elixir and incredibly surprised by how little I need static types. They are useful, but less needed now.
Never could believe I would say that. But I'm the most productive I ever been.
I get bugs of course, but they are related to queues lengths, retries, api errors, memory usage, performance, ... Bugs related to incompatible types are rare.
But also Elixir's paradigm is not the same as lets say some PHP slinging or JS. It encourages a stricter functional style, that already makes for better code than what many people produce in languages like JS and PHP, or Python.
Im not that familiar with elixir, but erlang is hands down the best dynamic language to build systems in. It feels like they made only right decisions. I can only assume that elixir has made improvements on that.
Meanwhile java and kotlin have the least useful types I've encountered in my career. Scala has some pretty powerful stuff, but it's like the c++ of powerful types. The ergonomics aren't great. Finally, typescript has a lot of impressive typing constructs but it really falls down on the runtime side (because javascript).
The real comparison to make would be erlang/elixir vs rust or ocaml.
Lots of modern (...3+) python code uses type hints and a type checker. It can be as strict as you'd like it to be, which is exactly how I like it. It's what pulled me away from ruby.
Meanwhile, static languages are too often a giant pain in the ass, and in return for writing a lot of annoying code, you get in return guarantees that only really apply within your process's memory. And in a microservices world... you're actually realistically using the protobuf type system. Which generates just fine for python. And then "internally" you can use python's type checking where it helps, and if it doesn't help, then for that bit of the code, simply don't use it (and write "true" python).
I also find that a HUGE problem in the world is that programmers just. can't. help. themselves. They LOVE to over-define. LOVE IT. It's a siren's song!! Static type systems are a trap for the part of our brain that loves to architect. One of my favorite things about python is that it helps programmers _let go_. Not everything needs to be an interface. It's python. Everything is already an interface. Now just write the code without all the distracting 20 layers of indirection. And if we ever need one more, it's python - it's practically already there. Just make a new type, put @property on some methods, and you're good.
Obviously there are times I'd not use python. I could foresee myself writing Rust if I had to do code where correctness was of utmost importance (like, crypto, or embedded software for a medical device where someone's ventilator is hooked up to it, or similar). But if nobody's going to die (so... medical and cryptography...) then I'm using python almost no matter what I'm doing. And I'll use numpy or write a C module if I actually end up needing true CPU-bound performance for something.
As much as I respect this guy who tried to work and push an alternative ecosystem, it's hard for me to shake off the impression that, rather than due to Haskell compile time, he moved to python because it's easier to find developers for it and it's the de facto scripting language for LLMs.
No problem about that, of course. Running a company is hard enough, I think that passion and idealism for a language/platform/technology out of aesthetic appreciation can only go so far and after a certain age just making money and reaching your professional objectives count more.
Fast compile times is one of the most important qualities for developer productivity. It made Haskell a non-starter for many developers even before LLM driven development took off.
> So far, we haven’t lost much in the switch. The type safety we gave up hasn’t been noticeable in any concrete way yet, especially considering our test coverage has never been better.
Moving to Python means you'll have quicker compile times but now you'll need a bigger test suite which will take longer to run to get feedback after code edits? In TypeScript, you can compile and execute without waiting on type checking, an option like that would help?
> That means taking AI seriously as a first-class user of the ecosystem.
Honestly for me any time a language or a tool markets itself as "for the AI era" or takes agents seriously as first class users, I run. It's a bad smell for me.
I'm happy that there are things like Haskell that are still focussed on correctness and sanity, and not pandering to the AI psychosis driven on by the market.
I am really disturbed by this ideological framing of "it's the AI era now" "we have to let the agents run" "speed, speed, speed" "if you don't learn to engineer with random garbage, you will be left behind!!"
Some of us will need to leave the cult of the empire of AI and live in caves like the desert fathers, committed to actually crafting correct things.
The smell follows you but atleast we don’t have to smell it any more. Thank you.
Also typescript is a better overall language than golang. More expressive types. Golang wins in performance and simplicity and raw ugliness.
Haskell is great, once you write code, chance of certain kind of bugs appearing is very low.
Biggest problem i've is i develop on Apple Silicon and can't cross compile to x84 linux which is most common deploy platform for servers
Compare this is to Go, which is what i use for almost everything else, cross compile, copy to server and there you go! So simple.
this causes a lot of pain and suffering and slow development loop in Haskell.
I'd use Haskell for more things if someone solves this issue.
Please haskell community, come up with a way to solve this issue. It wouldn't be a problem if i've to install a JVM like we've for Java.
- The benefits of more "extreme" type systems are more accessible and valuable than ever. I have a fairly involved project built on Lean that I hope to open source this month, and it's been a joy to work in even for uses outside of mathematics.
- Readability, build time, infra complexity, and everything that affects your speed after finishing your implementation--these things now matter more than ever.
It's sort of a dual ergonomics problem, in some sense. And given that, the author's lament makes complete sense to me, especially:
"An AI-enabled Haskell ecosystem would ask different questions. How do we make Haskell easier for agents to use well? How do we get more high-quality Haskell examples into model training data? How can we scale reviews? How do we make library docs full of copy-pastable, realistic examples, not just beautiful types? How do we make project bootstrap fast? How do we make error messages more agent-friendly? How do we reduce cold build times? How do we make common industrial patterns obvious to a model that is trying to help?"
[0]: https://discourse.haskell.org/t/after-7-years-in-production-...
I wonder if one of the reasons I've been getting such excellent results from LLMs with Clojure could be that they can immediately try things out in the running REPL. Rebuilding my entire application from scratch takes minutes, while connecting to a running system through nREPL and compiling single functions or namespaces takes less than a second.
> The model can often avoid the mistake before the compiler ever sees the code.
> The type safety we gave up hasn’t been noticeable in any concrete way yet [...]
> Type safety can be a huge advantage for LLM-generated code if the compiler is helping the agent converge quickly.
Well, good to have this question cleared up once and for all :-)
Edit: I dont quite buy that argument, in case it was not clear.
Edit2: oh, is it specifically about incremental compile time? In which case I understand that tests can be run piece-wise quick indeed.
Also I assume the cost of development skyrocketed so much over the months that our company started implementing drastic cost reductiom measures everywhere that seem to mitigate any improvement we've had.
That's where one better appreciate the work being done in Zig to get incremental compilation in milliseconds.
The choice of Python is quite interesting, and a big swing in the other direction.
___
Please don't let your kernel panic, I'm being sarcastic. You never know the level of emotional attachment some HN people have to their tooling...
"solver failed" is in theory a better error. it fails faster and is unambiguous. maybe because it's weird it is scarier? idk
the main issue is managing PVP bounds is a little hard to keep correct. but cabal also lets you "allow-newer" for duct tape.
It seems the job of a senior software engineer these days is to make fun of programming languages on HN pointing out their flaws, while simultaneously keeping an eye, waiting on agents doing "thinking".
i could probably do my job and get passing (or better) perf reviews using nothing but fundamental-mode and the compiler.
I find Ruby a very beautiful language, and Rails is an excellent web framework, but I need typed functions, record types and sum types.
They help not just with correctness, but also as living documentation that lets me understand AI generated code. TypeScript provides discriminated union, but not exhaustive pattern-matching, and its syntax is a bit verbose, but since I'm no longer writing most of the code myself, I can live with it.
However I can't imagine using Python or any other dynamic language going forward. There is likely good reason for you to choose it, and I'm curious to know what that is.
this is undoubtedly true.
We need more general purpose Elm languages in the space.
- haskell exceptions and laziness are devastating for production
- too small ecosystem, had to write 10+ SDKs (now with AI that's less of an issue)
- haskell ecosystem is too fragmented due to prima donna prevalence
I'm skeptical; I would think that the problem would actually be that mistakes in a large body of Haskell code would be difficult to fix: that massaging the generated slop into compiling ranges from unpleasantly time consuming to intractable.
Might the author be hiding the honest statement of the problem: that he would rather move fast and break things as a slop artist, but the guard rails are too rigid in Haskell?
Who cares about performance.
LLMs have made me move away more from python rather than into it. I'm very surprised by this experiences of the author. The article is all over the place as well. Going basically all in on Python because it is apparently better than Haskell for LLM use and than agreeing with someone that says Rust is the best.
Another obvious point is that an industry that runs on code slop will stagnate in terms of language an human tooling design.
Long compile times can be an issue, but it’s an issue that can be managed.
Would I scrap Haskell in favour of Python, now in the age of agentic coding?
No. No I would not.
It looks like the issues are in the compiler and documentation so hopefully it’s fixable… I write in Python every day but I do miss smarter languages and I hope AI doesn’t fully obliterate them.
With AI coding emerging, a single person can now churn out 100,000 or 200,000 lines. And realistically, from my experience, once you go past 40,000 lines, it's hard to memorize everything. So what do you do? Human coding shifts toward writing tests and gates, and once you feel comfortable that things are safe, you add more features.
AI coding takes this to an even more extreme level. Learning Haskell is great for learning domain modeling—I learned domain modeling through Haskell myself. But now that AI has become genuinely useful, it seems like a fundamental shift in workflow is happening.
Realistically, for commercial competitiveness, the domain of code black boxes is getting larger. The number of lines a single person has to manage is increasing, but their cognitive limits haven't changed. Even the amount of background knowledge required keeps growing.
In that sense, I agree with the author's point. It's not that Python is a better language than Haskell, so people switched—it's that Python has almost no build-up process and it serves as the standard interface for AI models.
The value of a language as a product doesn't come from the compiler's excellence. It comes from its users.
Avi Press's article ultimately reflects the reality that if you can't stay ahead throughout the entire lifecycle, you'll fall behind your competitors.
Maybe I'm just echoing my own thoughts, but it's reassuring to see that a well-known programmer thinks similarly to me.
It's interesting in particular because the argument of the article has at its core nothing to do with coding agents:
"so far, we haven’t lost much in the switch. The type safety we gave up hasn’t been noticeable in any concrete way yet, especially considering our test coverage has never been better."
people said the exact same thing when they moved from Haskell to Python or to JavaScript before the latest tech. Tests, tests, tests, and faster development cycles is just the language of the Agile people who have been advocating for this for decades. The people who didn't buy it never did so because the claims about development speed were wrong, they didn't buy it because they had a fundamentally different outlook about what matters in a codebase over years. I'm interested to see how this will look in three years rather than three weeks. If you're so seduced by the idea that shipping next months feature faster is so important I honestly don't know why you ever chose Haskell in the first place.
if my CEO wrote this article, I'd quit [1] in an instant
[1] "quietly" while i found gig+1. oh and the private out-of-band engineering gossip and trash talk would surely be hilarious
Clothes became chain-stitched (and later lock-stitched) because machines could chain-stitch. If there were a super-efficient hammering machine, it could be better to figure out ways to use nails to replace screws in designs than to hold onto screws just for nostalgia's sake.
How about for the sake of the bridge continuing to stand? Or is that not a good enough reason for the accountants?
Everything sort of exists, but it's this heinous zero documentation, high pain tolerance thing: buck2 and RBE with NativeLink and hooking that up to action runners and it needs to all work in a container or on nix or in a deb and on MacOS, you hand roll the auth and the certs and where do your compilers come from, can it do NVIDIA, can it do mobile.
Problem is switching off Haskell doesn't help for long: the agents proliferate and you're back where you started with more bugs. So I've been sucking it up and getting all this shit one click and it works. This is good enough for my use, and if the Scarf folks want a solution and are willing to work with a garage band startup, I'd be open to doing a closed alpha. I have a buck2 where you write the rules in Haskell (if you even need to change the prelude, it ships with a WASM that isn't coupled to fbcode), and the Nix cache/substitutor is backed by NativeLink so it scales to anything and it speaks all the protocols correctly and with a verified supply chain.
I'm not even really sure this will become a product, I just need it, but I sort of suspect others will need it too. If there's interest I'll put up a landing page with an email sign up thing.
Which it always has been, whether outsourcing to cheap "talent" overseas, or now a subscription service text generator.