Hacker News Reader

>If you have AI systems that can simply build out POCs in days, backtest on real data, show reliable results and numbers, you get a suite of product options you were never able to get before. If you have coding agents that can speed up implementation, you can build more stuff and choose the things that stick.

I'll also add this: within a large organization, you often need to interact with many different codebases owned by many different teams. Agents have made it much easier to wrangle by having the ability to deploy one to scope out your web of dependencies to learn about what would be needed for feature X, and how that integration can happen.

We've been doing far more away team work simply because it makes things move faster. It's easier to convince a team to sign off/review something than it is to get them to commit to the planning and eventual work.

It genuinely is helping things move faster inside large organizations. Or at least, it is for us, particularly since we're getting organizational prioritization to actually build the scaffolding to make those agents more effective at search.

> It's easier to convince a team to sign off/review something than it is to get them to commit to the planning and eventual work.

1000x yes: you have touched on what I think is the single biggest factor here, that is the humongous value of POCs. they are gnarly to build without agents, and so we used to have to get everyone on board so we didn't get screwed in performance reviews, which was monumental task because that means convincing very busy PMs who have a lot on their plate and dont want to take risks on things they don't understand, and now it's like "can we scale this out" and you have a very nicely formatted proposal and POC. It de-risks things very quickly

Pieces of concept and other prototypes have always been cheap (see hackatons). The main issue is that as soon as you’re touching customer data or modifying process they’ve paid you for, you have to be really careful. No one wants to be responsible for an outage that cost the company its biggest customer.

Yes, but at scaled companies, where building a simple POC hooked into real systems is nowhere close to easy. To the point where it means that you might as well just do the full thing. That's where the enterprise spend and the impact is.

Isn’t that a matter of configuration management? Or do you want to alter the real systems as well?

historically it's been a matter of an absolutely horrific amount of Kafka-esque problems.

Say I want to build a feature in a product.

- DS has to do a deep dive (need buy in) to opportunity size and derisk with data. That DS has to work with other DS (people may have left or moved teams) to figure out how to get the right data and figure out what the difference is between 10 different tables that have overlapping but inconsistent data. - Eng has to build up an actual simple demo (need buy in) - Design has to make it not hideous (need buy in) - Legal has to review what you're doing; POCs should involve real data where possible because otherwise no one will trust it, even if its just for user analysis on existing products

This plus about 6 internal system bugs for custom tools that are flaky and who's team has long been re-orged or laid off, 8 people who won't answer you, 2 PTO's for the stakeholders, 6 weekly meetings

no one did POCs, they just had ideas and tried to get PM's to put it on the roadmap so if it fell through at least it was bought into

Legal approvals won’t be in that category.

You still want someone whose ass is on the line if they get it wrong.

Absolutely but you want to package it to them nicely and efficiently. The biggest blocker is legal and everyone else speak two completely different languages and we often don’t know what’s important to flag and legal doesn’t know enough to ask all the right questions. Also, many things can be templated, and in an industry where regulations and precedents change so quickly, agents are at the very least a good tool to flag issues (e.g. we were approved to use data X for Y but now decision Z negates this). The propagation of this information is not very effective now and legal review at tech companies, while absolutely essential, is somehow a worse experience than going to the DMV when it’s crowded.