namanyayg

Seeing the suggestion influences the decision

You need to set up a methodical go-to-market process.

You can automate some of it, but it's good to start manually and learn the ropes first.

You need to understand who your ideal customer is and what channels are they on.

Linkedin is a common one for b2b sales.

If you make 200 connections on linkedin every week and send them a message, some % of that will convert to meetings and sales (the exact % depends on your audience, message, and how urgent the pain you're solving is)

It is extremely hard at start, especially for engineer types (like myself)

Fortunately, it's a skill like any other that can be learned.

Read this article for more info https://x.com/namanyayg/status/2097783972302557668?s=46

just make this a first-party thing https://agentkind.io/relay

OP is from 1998 so it's the other way around

I had no idea you could do something like this through the Pebble!

I pre-ordered mine last year but it looks like it's not going to come before September :(

As soon as it gets here I think this is the first project I'm going to attempt. It'll be helpful you've got any resources or source code for the Pebble STT and TTS parts. (I already have my own hermes that I want to hook it up to)

Exactly the same experience and I try to wake up with natural sunlight as much as possible.

The productivity and mental boost from that is incomparable.

I cannot disagree more strongly. In the correct setup, there is always a tangent line between all parts of the screen as you swivel your neck horizontally. I did not have that without an ultrawide and it led to various neck muscle issues that I regret

I should have added that I ran my first startup in India for 8 years before moving to the states, that's why I have an insider view.

One aspect that seems to be missing is the abuse of the free credits provided for new companies by AWS, Azure, and other providers.

I know of a friend's company in India who purchased inference, at 4% of the actual price and states that it gave him an unbeatable competitive edge in their large running video influence pipelines. Any new competitors could not offer their pricing at all.

Primarily that operated because registering a new company getting free AWS credits was a very tiny cost

I respect your opinion and experience, but FWIW for the readers, I had the completely opposite experience.

The batch did divide into groups. Some people were learning functional programming, others focused on Transformers architecture, while yet others on dynamic programming. But I quickly found a small group of people who were building things that I was interested in, games and AI applications.

In the unstructured time I was able to set my own goals, ship a lot, write a lot, and give a lot of talks on topics that are very interesting to me.

I didn't expect or get any free food from the Recurse Centre -- I was in New York, I had the world's cuisines available to me. I often went on lunch along with other RC members to reasonably affordable places within walking distance.

RC is not a college course with exams and a fixed syllabus -- It was unstructured and self-directed -- and that's what made it so special.

Thanks for making Recurse so awesome Nick, and it was great running into you at demo day.

I learned so much and made great friends at my six weeks in Brooklyn.

Without Recurse Center I probably wouldn't have made it into YC with a startup of my own today. Long live RC!

Nice, I built something similar ~5 years ago using nested git repos and scripts to accomplish a similar purpose of combined private and public repos.

My shell script definitely wasn't google scale tho!

This was one of my favorite Project Euler problems, but getting to the mathematical solution admittedly took me a couple of weeks.

Perhaps because I was pigeon-holing this as a programming optimization problem.

I wrote about it too! [0]

[0] https://nmn.gl/blog/vibe-coding-gambling

Amazing and industry-changing. Is he able to update it by himself now?

I have a similar story with my washing machine repair. I went through 2 service technicians not being able to diagnose it. Gpt did it and I told the 3rd what to do and it worked.

Yup it's gotten good enough that it mostly works (thats why i built my startup around this idea)

You can live in a house, but you can't eat stocks

One of the linked demos, "rainbow surf", got me hypnotized. https://www.youtube.com/watch?v=QKLhH_ANwIc

very true! we sandbox and keep it independent of the main codebase, which makes it easier to build and share new features.

Nice article. That's basically what we do. We also make the created code saveable and shareable so they can reuse what they've already built

Yup! That's one of the biggest things we build while shipping this for a Series B SaaS. It's invisible but one of the hardest technical challenges.

We've got guardrails before any code is generated and we put three layers of validation after the feature is ready to make sure it's using the correct data and it's functioning for the use case that the customer asks for.

Yep, even our use cases are 80% read-only, and the rest where we write is very constrained and guardrailed.

Giving non-technical users the freedom to create new backend objects is definitely scary and not worth it imo!

What did you build, any links?

Our solution to that age old problem is to scope each new feature into its independent "app", which is sandboxed and walled off from the main source code

Thanks for the new tagline ;)

Manual qa, evals, prompting, and inhouse validation steps. AI is not 100% perfect but it's working well for most use cases.

Yup 80% of the usage is for new frontend features and that's actually quite valuable for our customers who serve traditional enterprises like manufacturers, hotels, labs.

Only some (very well guardrailed) backend functionality is being used.

We're live with five businesses and 2000 daily users! The technical challenges in setting the mapping, ai model, and dsl that creates reliable outputs is the engineering value we deliver.

Yes -- the future of SaaS looks more like a system of record, on top of which AI serve what's needed on the fly based on what the customer needs.

Exactly ;) One of our customers say "we've given their entire team superpowers"

I've been dealing with technical debt for half of my career. Here's what we're doing to prevent it:

- We don't ship to prod or to the main repo -- each feature is a scoped, sandboxed, separately version controlled "app".

- We have a proxy layer to pin API versions, so if the underlying contracts change, we still support all past created apps.

- Authentication follows your SaaS platform's RBAC and authentication tokens, making it easy to share within a team or across multiple tenants.

Looks like the author on this just copied from chatgpt. Who gave this drivel 6 upvotes?