nicodjimenez

I actually just built an ATS for my company Mathpix. But it never occurred to me to use resumes. Basically we have a set of company values and a specific open ended questionnaire to gauge the fit:

https://mathpix.com/careers/apply

Then internally we have dashboards and sorting based on AI agent scoring. I noticed the scoring is imperfect but still saves a lot of time. Candidates scored at or below 2/5 are reliably bad and candidates above 4/5 are consistently impressive and leave thoughtful answers.

The biggest thing is not using resumes. You can’t reliably gage applicants without a writing sample and resumes are the worst form of writing sample. Also you need to be intentional about who you’re hiring for, both to craft the questions as well as grade the responses.

tokenmaxxing is silly, but if a developer or manager NEVER uses AI then I do think that's cause for concern as it shows a genuine lack of curiosity... perhaps tokenflooring makes more sense than tokenmaxxing

Part of what makes Hendrix's live performances so great is how completely unreproducible they are. Even Jimi himself could never recreate that one note sustain when he begins the solo on Machine Gun. To re-create it, you'd have to set the room up exactly the same, tune the guitar exactly the same, position the guitar relative to amps exactly the same, etc. So Hendrix being very sensitive and connected to the room was able to harness that energy into something unique that stands the test of time. Machine Gun is well known, but his Red House performance at Randall's Island also stands out to me as exceptional, those are the 2 key Hendrix performances. I read somewhere that Miles Davis was really impressed by Machine Gun and you can see why.

One thing I learned after buying some gear at home to try to record electric guitar at low volume is how important the physics of the speakers are. You can plug a tube amp into a cabinet simulator and you'll lose a lot more than using solid state electronics on a good but not great Fender amp, especially if you use fuzz / distortion pedals.

I'm not sure Hendrix was a systems engineer, but he was a transcendent blues artist, that's for sure.

great article and many great points here

What I dislike about Claude code and vibe coding in general is that I haven’t seen Claude code users learn a whole lot about how to do their jobs better. A terminal pane is just too small to be a good place to learn.

With vibe coding you just give the code some constraints and then system will try to work within those constraints, but what if those constraints are wrong? What if you’re asking the wrong question? Then you’ll end up with over complicated slop.

It’s a shame that vibe coded slop seems to be a new standard, when in fact you can use AI tools to produce much higher quality code if you actually care to engage in thoughtful conversations with the AIs and take a growth mindset.

My main take on microservices at this point is that you only want microservices to isolate failure modes and for independent scaling. Most IO bound logic can live in a single monolith.

This seems completely wrong. In an RPC call you have a trivial loop, for example.

It would make more sense to say that the event tree should not have any cycles, but anyway this seems like a silly point to make.

Microservices in Golang are definitely related to classes due to the ergonomic aspects of a language. It takes a lot of discipline in Golang not to end up with huge flat functions. Golang services are easier to reason about when they are small due to the lack of abstractions, also Golang is very quick to compile, so its natural to just add services to extend functionality. Code re-use is just a lot of work in Golang. Golang is not monolith friendly IMO.

The solution to bad abstractions it not to make it very difficult to create abstractions at all. For systems code I think it's fine but for application code you probably want some abstractions or else it's very hard to scale a codebase.

Golang to me is a great runtime and very poor language. I could maybe get used to the C pointer-like syntax and to half of my code checking if err != nil, but the lack of classes is a step too far. The Golang idiomatic approach is to have a sprawling set of microservices talking to each other over the network, to manage complexity instead of having classes. This makes sense for things like systems agents (eg K8) but doesn't make sense for most applications because it complicates the development experience unnecessarily and monoliths are also easier to debug.

I would not use Golang for a big codebase with lots of business logic. Golang has not made a dent in Java usage at big companies, no large company is going to try replacing their Java codebases with Golang because there's no benefit, Java is almost as fast as Golang and has classes and actually has a richer set of concurrency primitives.

Note that Mathpix Snip can quickly convert such screen shots to markdown code via keyboard shortcut. Disclaimer: I’m the founder.

There are only two political systems at the end of the day: authoritarianism (everyone knows who is in charge) and oligarchy. Populism, in a liberal democracy, is basically authoritarianism-lite representing the interests of a particular faction of oligarchs. There's no "populism" in China, that's an American & European invention. Populism is ugly but it's a useful tool that can get things done in an oligarchy.

What kind of codebases do you work on if you don't mind me asking?

I've found a huge boost from using AI to deal with APIs (databases, k8s, aws, ...) but less so on large codebases that needed conceptual improvements. But at worst, i'm getting more than 10% benefit, just cause the AI's can read files so quickly and answer questions and propose reasonable ideas.

The most important piece ever written about startups, probably. Applicable to doing anything new.

For startups, the devil's in the details though. The goal is to scale but you get there by doing things that don't scale successively.

For accurate and easy PDF to Markdown / LaTeX / JSON check out:

https://github.com/mathpix/mpxpy

Disclaimer: I'm the founder. Reducto does cool stuff on post processing (and other input formats), but some people have told me Mathpix is better at just getting data out of PDFs accurately.

I tend to agree with this. These days I usually use LLMs to learn about something new or to help me generate client code for common APIs (especially boto3 these days). I tried Windsurf to help me make basic changes to my docker compose files, but when it couldn't even do that correctly, I lost a little enthusiasm. I'm sure it can build a working prototype of a small web app but that's not enough for me.

For me LLMs are a game changer for devops (API knowledge is way less important now that it's even been) but I'm still doing copy pasting from ChatGPT, however primitive it may seem.

Fundamentally I don't think it's a good idea to outsource your thinking to a bot unless it's truly better than you at long term decision making. If you're still the decision maker, then you probably want to make the final call as to what the interfaces should look like. I've definitely had good experiences carefully defining object oriented interfaces (eg for interfacing with AWS) and having LLMs fill in the implementation details but I'm not sure that's "vibe coding" per se.

Check out mathpix.com. We handle complex tables, complex math, diagrams, rotated tables, and much more, extremely accurately.

Disclaimer: I'm the founder.

Thanks for the kind words. What are some of the annoying issues?

You can get bounding boxes from our pdf api at Mathpix.com

Disclaimer, I’m the founder

Check out mathpix.com we have a hybrid approach towards OCR that features accurate layout understanding (with accurate bounding boxes) plus accurate OCR outputs.

Disclaimer: I'm the founder and CEO.

Shameless request for help: if anybody has experience with seq2seq on TPU, and you want to do a cool project to deploy a world class Pytorch image parsing model to TPU (and do this quickly), please contact me immediately for a well paid and interesting job opportunity at nico [at] mathpix.com.

Agreed, something simpler than Nomad as well hopefully.

https://mathpix.com/careers

Not in baseball.

I recently watched some Mets games and it’s just incredible how commonly you use a left handed hitter go up against a left handed pitcher.

Have you checked out Mathpix? It's another option.

Disclaimer: I'm the founder.

I bet that Reed Hastings has a sub organization at Netflix that is run in founder mode, even if most of the organization is run in manager mode.

Whether orgs are run in manager mode or founder mode depends on whether there is a founder level leader available and nature of the changes that need to occur for the organization to remain competitive. Some orgs or sub-orgs cannot afford to have a founder making radical changes, because the risk this will lead to an exponential rise in defects for end customers is greater than the potential benefit.

PG tailors to startups and for startups the risk of the wrong product is generally much greater than the risk of product defects. So I tend to agree with his points here.

Exciting to see more work in this area.

See https://www.plasticlist.org/ for a recent SV based project to measure microplastic contamination across popular food and drink items.

You’re right :) got to love HN, I broke the code and got downvoted! I like a community that has teeth

Mathpix is going to lower price a lot for larger scale pdf processing, stay tuned!

Second page is even worse. Ends in repeated \cdots and doesn’t finish parsing page. Also it read number 73 as 3 I guess because the previous section number was 2.