Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why AI Infrastructure Startups Are Insanely Hard to Build

nextword.substack.com

Post reply on HN
Print view View original

131–140 of 190 posts

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#131

Post by nextworddev »

cootsnuck wrote: post #106

I work for a foundational AI company. I guess we're technically AI infra. We're inherently "narrowly" focused since our origin (which was well before the recent hype in past 2-3 years). Our customers are really the type of AI infra companies being talked about in this article. And yea, the new ones I work with everyday are often a dime a dozen. A revolving door of small startups trying to make the same general purpos…

Author here. Thanks for the perspective. p.s. I do hope AI startups not estimate how hard it is to break into vertical markets which have their own challenges

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#132

Post by staticman2 »

gizmo wrote: post #125
amarcheschi wrote: post #90

Earlier quoted context omitted.

Does their return trail s&p 500? I guess there are some of them that are successful and some that aren't, but to be honest I've only heard about vc as people who on average have a very very high return. Now, it is very possible that's just survivor bias, but I would need to search for some data

The S&P 500 is a list of the largest and most profitable public companies. It's hard to do better than the most successful businesses. Most other indices and hedge funds don't outperform the S&P 500. Most private equity shops don't. Most real estate investors don't. So it shouldn't come as a surprise that venture capital doesn't. Most startups don't get big. How many startups founded in the past decade have become hu…

It's hard to do better than the most successful businesses is not a statement that makes sense from the investor's perspective.

The price of an investment is based on the expected profitability of a company, an investment in a barely profitable company, if priced correctly, should yield returns at least equal to good companies like Apple, Google, and Microsoft, as the investment would be discounted to compensate for the poor expected future earnings of the company you are investing in.

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#133

Post by rapsey »

criddell wrote: post #130
rapsey wrote: post #24

Earlier quoted context omitted.

> Even OpenAI doesn't really have a product. They are making a ton of money off subscriptions.

I'm skeptical that $20 / month is enough to run OpenAI and be profitable. I would bet the real number is an order of magnitude higher.

I did not say profitable

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#134

Post by tonetegeatinst »

brummm wrote: post #120
ankit219 wrote: post #67

Great article. I am not going to name names, but over the last one year, whenever there is a concept that became popular in Gen AI, thousands of startups pivoted to doing that. Many come from software background where the expectation was that if the code works on one dataset, it would work for everything. You can see this with 1/ Prompt engineering 2/ RAG 3/ and now, after Apple's WWDC, it's adapters. Enterprises I h…

Random question, but what are Apple adapters? Kind of hard to google it, lol

Probably the name of the way you had the think differently to charge the desktop mouse upside dow.....

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#135

Post by BobbyJo »

ghaff wrote: post #123
BobbyJo wrote: post #100

Earlier quoted context omitted.

> Want to solve a real problem, help me create custom benchmarks, clean my data, get my small parameter model to reason better etc. I recently started a company with a friend if mine to do exactly this. Ive worked at a few AI startups over the last 8 years, and the problem everyone tackles independently (and poorly) is the long tail of dealing with input data that isn't great. You build a demo with sample data that w…

Ah. Cleaning data. The big trick going back to data warehousing.

When you know the underlying consumer is an AI model, you can do a lot more to make the input data useful.

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#136

Post by hibikir »

rgavuliak wrote: post #116

It's as if all of the AI devex/infra companies are cargo culting the story of how the people that made the most money in the gold rush were the people who sold the tools. The thing is that the tools were well understood and battle tested.

And this isn't just a matter of AI: You see all kinds of companies trying to provide value adds on top of cloud: "We will annotate DNA for you as a service!" When all they do is dockerize the same tools their customers use, and serve as small shims for the least sophisticated customers. The moment said customer grows, they understand they can replace the vendor with less than a week of work.

The people making shovels make the money by having strong profit margins, becoming a default vendor, and having a moat. Good luck doing that in AI!

And my favorite counter example of selling tools is precisely docker: They built tech used everywhere... yet how much value they captured? It's tge same story all over dev tool space.

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#137

Post by p1esk »

criddell wrote: post #128
brummm wrote: post #120

Earlier quoted context omitted.

Random question, but what are Apple adapters? Kind of hard to google it, lol

I know almost nothing about this stuff, but what I know about Apple adapters I learned from this page: https://machinelearning.apple.com/research/introducing-apple... > Our foundation models are fine-tuned for users’ everyday activities, and can dynamically specialize themselves on-the-fly for the task at hand. We utilize adapters, small neural network modules that can be plugged into various layers of the pre-traine…

Sounds like regular lora adapters

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#138

Post by TimPC »

rapsey wrote: post #79
pas wrote: post #68

Earlier quoted context omitted.

if they are not profitable with these prices ... what the fuck they are doing!? do they just have company coke-athons all day every day?

It is a sales and engineering heavy business. Very difficult to generate large returns.

They are becoming like SAP though where initially a company buys one service and it soon finds itself buying every adjacent service from them. If they manage to do that successfully they will be quite profitable.

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#139

Post by ankit219 »

brummm wrote: post #120
ankit219 wrote: post #67

Great article. I am not going to name names, but over the last one year, whenever there is a concept that became popular in Gen AI, thousands of startups pivoted to doing that. Many come from software background where the expectation was that if the code works on one dataset, it would work for everything. You can see this with 1/ Prompt engineering 2/ RAG 3/ and now, after Apple's WWDC, it's adapters. Enterprises I h…

Random question, but what are Apple adapters? Kind of hard to google it, lol

Sorry for being vague. I meant LoRA, but used Apple as an example because their demo showed the potential. At a conceptual level, you can finetune a base model to be good at a specific task - eg: summarization, proofreading, generation etc. These finetuned weights are at the top layer and can be replaced by other weights for a different task as needed. Apple demoed different tasks by showcasing how their model identifies the task and then chooses the right set of finetuned weights. Apple called it Adapters as it comes via LoRA (Low Rank Adapters). It's around for some time, but only shot into prominence after people got some idea on how to use it.

Re: Why AI Infrastructure Startups Are Insanely Hard to Build

#140

Post by datadrivenangel »

pas wrote: post #68
rapsey wrote: post #30

Earlier quoted context omitted.

> Who knew that Snowflake and Databricks would emerge against the incumbents. Snowflake is not profitable. I doubt Databricks is. Their market and business is crap.

if they are not profitable with these prices ... what the fuck they are doing!? do they just have company coke-athons all day every day?

They have a lot of sales people and sales engineers.
Post reply on HN