HNHacker News
TopNewBestAskShowJobs

sbpayne

186 karma · joined April 24, 2015

skylarbpayne.com me@skylarbpayne.com
submissionscomments
sbpayne··on If Dspy is so great, why isn't anyone using it?
yeah what I'm trying to get across here is that: Dspy does not solve an immediate problem, which is why many feel this way and consequently why it doesn't have great adoption!

But on the other hand, I think people unintentionally end up re-implementing a lot of Dspy.

sbpayne··on If Dspy is so great, why isn't anyone using it?
hahaha this has just been my entire last few years of experience :)
sbpayne··on If DSPy is so great, why isn't anyone using it?
This is a surprising take to me! Would love to learn more about what you mean. I feel like the problems they solve seem so direct to me. For example: RLMs are an approach to long context problems. Not every problem is a good fit for RLMs for sure, but I can see some problems where I imagine it would work well!
sbpayne··on If DSPy is so great, why isn't anyone using it?
Yeah, I think Dspy often does not really show it's benefit until you have a good 'automated metric', which can be difficult to get to.

I think the unfortunate part is: the way it encourages you to structure your code is good for other reasons that might not be an 'acute' pain. And over time, it seems inevitable you'll end up building something that looks like it.

sbpayne··on If Dspy is so great, why isn't anyone using it?
I think this misunderstands the need for iteration! Maybe I could have written it more clearly :).

The reality is that you don't want to re-deploy for every prompt change, especially early on. You want to get a really tight feedback loop. If prompt change requires a re-deploy, that is usually too slow. You don't have to use a database to solve this, but it's pretty common to see in my experience.

sbpayne··on If DSPy is so great, why isn't anyone using it?
Oops! That's actually out of date from prior template I had. I don't actually consult at the moment :). Removing!
sbpayne··on If DSPy is so great, why isn't anyone using it?
I don't know that you misunderstood. This is one of my biggest gripes with Dspy as well. I think it takes the "prompt is a parameter" concept a bit too far.

I highly recommend checking out this community plugin from Maxime, it helps "bridge the gap": https://github.com/dspy-community/dspy-template-adapter

sbpayne··on If Dspy is so great, why isn't anyone using it?
I consistently hear great things from Dspy users. At the same time, it feels like adoption is always low.

Stranger still: it seems like every company I have worked with ends up building a half-baked version of Dspy.

sbpayne··on Show HN: Mirascope – The LLM Anti-Framework
This has been one of my favorite frameworks to use for AI development. I feel like it is pretty dead-simple to use. But also has a great quality: it gets the f** out of your way when you need it to. I feel like with a fast moving layer beneath it (ie ai providers), I need to be able to adopt and experiment with new features from the providers, without having to wait for your library to update.

I think the api generally is great taste. Having typed functions as the core abstraction (which shares some similarity with Dspy Signatures) was a great move.

Congrats on the launch, can't wait to see the platform come together more and more.

sbpayne··on Top Secret: Automatically filter sensitive information
This is great but it does not “prevent”; it reduces the chances of. NER is not 100% performant. It is very good in many cases, but use with caution!
sbpayne··on Why RAG Is (Still) Not Dead
As new model releases support longer and longer context windows, there is a lot of discussion around whether RAG is still relevant.

RAG is here to stay for a while:

(1) Enterprises have much more data than reasonably will fit in a context window any time soon (2) Even if you can technically put 1M tokens in, that does not mean the model can effectively use it all (3) Longer input = higher latency and cost for inference

Would love any other thoughts on the topic!

sbpayne··on Bauplan – Git-for-data pipelines on object storage
I have really enjoyed the conversations I have had with Jacopo and Ciro over the years. They have really revisited a lot of assumptions behind commonly used tools/infrastructure in the data space and build something that really has a much better developer experience.

So excited to see them take this step!

sbpayne··on Show HN: Simple to build MCP servers that easily connect with custom LLM calls
This is great! I personally found the original Anthropic MCP documentation pretty lacking in terms of how Claude Desktop used the MCP server(s), constraints, etc. For example, there is a pretty hard timeout which will cause the MCP server to crash.

Glad there's a simple to use solution for creating my own server where I can make some different design choices!

sbpayne··on Ask HN: Songwriters, what software do you use?
Basic features:

Create projects Projects can contain notes and audio (uploaded or recorded in browser) Then theres an AI chat in the project where the docs/audio are available as context (multimodal models used)

Its definitely very early; AI and UX need a lot of work. But definitely has helped me get over some “humps” with writing songs.

For extra context: I write songs with acoustic guitar and vocals, but I would say they are pretty simple overall.

sbpayne··on Ask HN: Songwriters, what software do you use?
I have been building my own app for this recently.

Its very early, but I have been shaping 3 songs with it already and am starting to get some friends to try it.

I am self taught with songwriting/music so I think it might reflect my own idiosyncratic songwriting process more than anything else at the moment.

Happy to open up a preview if anyone is interested though.

Shoot me an email if interested (in profile)

sbpayne··on How to stop fighting with your AI pair programmer
Interesting! I find I very often have to do this with humans. But yes, the repeat mistakes is a thing. This is where ensuring I output context along the way is helpful. Although it's maybe annoying to "re-refer" to the context, I find that it helps the AI make the same mistakes less often
sbpayne··on How to stop fighting with your AI pair programmer
I think there's some updates I can make to this to give some guidance around the (1) you have to know a lot of things point. You can use AI assistance to help learn more. But ultimately, I don't believe we are at the point where the AI makes you "smarter"; but it _can_ make you more productive.

I disagree with (3) based on my experience. That feeling happens when I am not providing enough context. I rarely have experiences where I step back to provide more context and still end up in a dumb loop. Highly recommend providing lots of context + breaking down the problem more.

I would love to dig deeper into an example you have where you feel you "never get your time back". Because in general I am saving a lot of time from how much less typing I have to do.

sbpayne··on How to stop fighting with your AI pair programmer
Interesting! Do you have an example with code? I am struggling to adapt the cheeseburger example to something with code.
sbpayne··on How to stop saying 'Fuck you Cursor'
I noticed a lot of people complaining about Cursor agent being bad. And my initial experience was bad. After working with it for a while, I found a workflow that works for me; I hope it's helpful for you too!
sbpayne··on On Building Git for Lawyers
It's curious to me how many people think "just convert it to a different filetype" will solve the problem.

Do you think there are other professions/industries that would benefit from this?

sbpayne··on Ask HN: Meal Planning App?
I started using an app called Intent -- it seems to support a lot of these features. I've been very impressed with it.
sbpayne··on Google's infamous internal 2010 “I just want to serve 5TB” video now public
I'm so glad I can see this again. I forgot how much I missed this.
sbpayne··on Do patents kill innovation? The US patent office is asking
For AI/ML related patents, I’m not sure innovation is stifled insofar as people mostly don’t pay attention to them.

Most of the techniques that _every_ company is using are patented, but no one seems to notice.

Admittedly I’m sure there are cases where it did stifle innovation in this space, but I think on average perhaps not.

It’s interesting to see the comments on eink patents — I was never aware of how much it stifled innovation in that space!

sbpayne··on Ask HN: Words of encouragement for someone lost in life?
I’m sorry to hear that you’re feeling like an outcast.

First, I want you to know these feelings are reasonable and you’re not alone in feeling that way.

Second, while it may seem hopeless or insurmountable, you can make changes.

You have an opportunity now in front of you. You have identified an area in life where you feel inadequate. But you don’t have to feel that way permanently. You can make changes to build up your experience in that area. And I totally understand that it feels hard to know where to start, how to make progress, or whether it’s possible. It can feel demoralizing. But if you’re up for the challenge, you can grow.

I have often felt this way and I’ve slowly improved over time — now I’m 28 and have done a number of things I saw everyone else doing when I was 22 feeling like I had already thrown away my life.

If you’re open to chatting about it and think it would be helpful, shoot me an email: skylar.b.payne (at) gmail.com.

Wishing you peace

sbpayne··on “Location-Based Pay” – Who are we to complain?
If the company is not “remote first,” I think there’s a reasonable argument that your value is not location independent.

That being said, I can’t really think of a good argument for anything beyond pay based on in-office vs remote.

sbpayne··on We don't need data scientists, we need data engineers
+1. I generally say that data scientists bring an amazing skillet to the table, but companies can only leverage 10% of it.
sbpayne··on How to Stop Endless Discussions
In my experience (mostly large tech cos), getting anyone to pay attention to your design doc/RFC (regardless of quality or length) was fairly difficult.

Its been a consistent complaint of many engineers at every company I’ve worked at.

sbpayne··on Why software engineering processes and tools don’t work for machine learning
I often see comparisons of traditional software engineering vs machine learning like this where the former is a linear sequence of stages without cycles and the latter is some kind of tight iterative loop.

But my work in both building backend services and building machine learning models has always looked like how the author describes machine learning models.

In my experience the "process" generally isn't that different; but the cycle times are long enough for machine learning that it can feel different.

sbpayne··on CEO of Cyber Fraud Startup NS8 Arrested by FBI, Facing Fraud Charges
Instead of fitting the product to the market, he tried fitting the market to the product
sbpayne··on Secret gyms and the economics of prohibition
I think this is closer, but perhaps needs a slight change. I don't think it's out of spite; it's not like people are taking _more_ risk AFAICT. So maybe something like:

"there's a wildfire and firefighters asked you to limit water use to only drinking water, but some people decided to keep taking long showers, washing dishes, etc"

← PreviousPage 2 of 3Next →