Introducing Gradient Ventures | Hacker News Reader

Introducing Gradient Ventures

gradient.google

125 pointsframschwartz28 comments

IMO this has significant appeal. Lots of firms can provide capital, but few can provide relevant expertise let alone training.

They seem to be trying to cover all bases with the Google Brain Residency the Machine Learning Ninja program, standard VC funding, and now this. If you have talent in ML/AI there is a way Google can help you succeed in the style of your choice. Want to be a founder? Excellent! Want to be a founder but also kinda part of Google? Sure! Are you super talented and experienced in other disciplines and want to explore AI and maybe contribute a 2-5% improvement to one of our model's performance? Yes! We have that!

Actually a 2-5% improvement over state-of-the-art is a research contribution in its own right.

In some problems this is incredible.

Google (and other large tech companies that have invested in AI internally) are probably the most valuable investors, because they have access, and can purchase, the largest and best data sets available. It's hard for AI start-ups to do anything without access to the right data sets, and large companies can and have better access to that data.

I've been meaning to create a marketplace for data to solve this problem. The obvious issue is how to protect sensitive information (perhaps instead of selling cleaned data, the platform trains models for you on leased data). I think the problem will grow quickly as meaningful data and the tech that it enables becomes an ever growing barrier to entry against startups and slow moving enterprises alike.

Check out data.world

> they have access, and can purchase, the largest and best data sets available

Google might have an advantage in personal data, that can be used for advertising and health, but when it comes to general data, such as image datasets and NLP datasets, they can be found in the public domain and are growing fast. There is just a specific, limited advantage to Google in datasets. Mostly for ads.

The best data set will in general only be as good as the raw data that was used to prepare it.

I think you underestimate just how far along Google is with respect to the huge amounts of raw data they handle. They've been around for 20 years now and amassed a lot of expertise handling all kinds of data imaginable at scale.

If you disagree, who would you say is ahead of Google wrt general data sets that are valuable?

The largest, most interesting recent public datasets in image and NLP were released by Google.

For example, here are some of their recent NLP datasets: https://github.com/google-research-datasets

In images, OpenImages is theirs, and there are assorted ones derived from YouTube.

Stanford's SNLI is the most recent non-Google NLP dataset which is getting used a lot. Babi (from FB) too, if you count that as NLP

That's true, but Google can also afford to acquire and monopolize data that other companies are sitting on but don't have the resources or talent to utilize internally.

>. It's hard for AI start-ups to do anything without access to the right data sets,

Exactly. As a founder of an AI focused startup, it is so hard convincing VC's that data can be as valuable as revenue.

Some VC's don't even bother beyond the screening call If there is less revenue although the data we gather in the process is more valuable.

Google very well knows the true value of data and hopefully they can shake the VC world for AIs

This is exactly why this kind of stuff should be prohibited, and the datasets should be legally regulated.

This is not just an issue for startups, but also for independent researchers at universities, who often can’t even replicate the successes Google and co report.

Most studies currently done in AI by Google, Amazon, etc were never replicated, and likely never will be able to, because access to data is missing.

Can you give some examples of such studies?

As a startup you have less leverage to access proprietary datasets. In fact, most of the companies we tried to work with wouldn't let us even access the data, let alone use it, unless we did everything on prem. Oh and one company even wanted to own half of our IP.

Your algos, pipelines, UI and UX might be the best but you still don't have the data to make it worthwhile on its own. If the dataset is really that good then giving a 50% stake (not necessarily the IP) might be very well worth it. It just comes down to the terms of the deal.

there's almost no scenario I can think of where giving up 50% of your IP is worth it. there would be almost no way to raise vc funds to hire more researches or devs.

I thought this was odd, from the about page

> We can help you find and incorporate data sets into your first models. From cleaning data to extracting the most important features, our team can help you get your production models to market.

While realizing the hardest part of a startup is everything but the tech, it seems odd they're telling AI companies they'll help with the hardest parts of the technical side, the ones that need to be done right well before anyone can tell if your tech has any merit.

I'd hate to be a first-pass reviewer for all the pitches they're gonna get. "I have this amazing idea, I just need someone else to build the AI behind it!"

Honestly, if the answer to that is that they invest in you and you use their money to buy google cloud ml services, and you go out and find customers, that doesn't sound bad.

I guess a big question here is, what does "AI company" mean to Gradient Ventures?

If it's "yet another company spewing buzzwords all over", ML-as-a-service might work, but I'm hoping it's more "adapting cutting edge AI research to do things people will pay for", which requires a hell of a lot more engineering and can't be done by just shoving their data into someone else's black box.

there is nothing that requires you to use GCP, Google ML services or anything along those lines. This was a surprise but true.

source: i am part of the Algorithmia team.

Right, I just meant in response to GP that said they didn't want buisinesses that wanted/needed much of their ML built for them. Google might be okay with that too.

Google is trying to find users for their cloud. A hot startup using GCP is good PR.

Really scary to see Google's maneuvering for total control of AI tech.

So, admitting my bias on this (I do AI at Google part-time), I don't see how GV is "total control". It's kind of the opposite - it's profiting from helping transition that expertise to new startups that aren't Google.

https://cogniac.co/ The demo over here seems to be fabricated. Even if you provide wrong labels interactively, the performance of the classifier keeps increasing ...

Picked up from https://gradient.google/portfolio/

I don't think they are trying to pass it off as a real demo.

Extremely impressed with the investment partners behind this fund. I worked with Shabih during my time at Google and saw his work KPCB; he's a deep thinker and also relentless advocate/supporter for the companies he backs. Anna Peterson is also a legend at Google. Think they'll end up being the strategic VC of choice behind every major AI company.

Reply on news.ycombinator.com