Like Airbnb, but for algorithms? | Hacker News Reader

Like Airbnb, but for algorithms?

bostonglobe.com

45 pointsprostoalex27 comments

Jesus this is one of the most contrived "Like Airbnb[or Uber], but for XYZ" similes I've seen in awhile...

> Algorithms are common, but also difficult to create. Aspiring entrepreneurs can take an eight-week crash course and learn how to program an app, but writing algorithms is a specialized skill. Algorithms are usually taught in higher-level computer science classes, and require a strong mathematical background...

Good algorithms can be hard to create, but anyone can create a algorithm, as it's just a recipe of mechanical steps. Anyone can create a "recipe" (of the cooking variety) for that matter.

Looking at Algorithmia's site, it basically looks like an eLance for writing things like scraper scripts, with the additional service of running it on their servers. Could be useful if their servers aren't EC2 and you need to run something that makes requests to a site that has blocked EC2.

We run on multiple public clouds. We do have a lot of scraper algorithms (mostly because the community created / wanted them) but also text analysis tools, computer vision , audio and video processing etc. By operationalizing these algorithms/functions/recipes we allow piecing together sacalable services for larger applications . Google has been doing this internally for years .

I don't mean to bash, but when you have resources like StackOverflow/StackExchange, Google Groups, and other forums that have most questions and answers for all sorts of software topics, and eLance and oDesk for hiring people to make whatever it is you need, what would make someone want to use Algorithmia?

I think there's a sizeable market for tailored code solutions that "volunteers" wouldn't help out with. The bounty model can be more efficient / less cumbersome than Elance for a set of problems.

A lot of people had this comment about Bountify.co; check it out to see what kinds of code bounties people post.

Good algorithms always end up as open source or as a scientific publication. Therefore I'm not sure if this business model is viable.

i think it's valuable in making the more easily accessible.

imagine a web dev who needs facial recognition but isn't savvy enough to put it together by hand?

having a simple API solves so many problems quickly

Even if the core of the algorithm is released, there can still exist better proprietary implementations. Google Pagerank, Random Forests, ROOT/SIFT, etc are examples of 'best in field' algorithms that aren't available for free, public use.

Pagerank is just an eigenvector calculation, and for a sparse graph it's linear in |V|+|E|. It's pretty easy to implement, I don't know what kinds of improvements you'd want. There's a ton of random forest implementations out there, what's the closed one that's so much better?

Also rootSIFT is patented so good luck with that.

Google has all sorts of stuff built on top of PageRank for their search engine - it's probably safe to say that their implementation is thus 'better' than the released algorithm.

One of the RF creators consults with a company (http://www.salford-systems.com/products/randomforests) on, I suppose, better implementations. Some people have claimed it's better than the R implementation, I don't know.

RootSIFT being patented is kind of my point - the authors created a (good) algorithm and haven't released it for free use. It's within these kinds of niches that Algorithmia can succeed.

I think there's some confusion. Pagerank is the leading eigenvector of the graph (G+1x) where G is an input graph and x is a weighting term.

This is not at all the same as modern Google ranking techniques, such as Panda, Penguin, and Hummingbird.

This has to be one of the worst analogies I've ever seen. This business model is nothing like Airbnb. It's more closely aligned with something like TemplateMonster, eLance, or oDesk.

I see no way you can rent out a algorithm for a business use and lease basis..which is what Airbnb is.

Other then that this I see the reason that something like this exists. Creating accurate mathematical formulas is definitely an art in itself that not just anyone can do.

Uh, so is this going to be able to host the massive database used to store the training data/model for some really large deep neural network or latent variable model? Because it seems that's most of what the "algorithms" on this service would be - just some application of machine learning.

No one is going to call this in order to run something like depth first search.

We do have a full data api for storing those models if that's what you want to do. We also help you train them , collect the data ,etc.

You can check out our blog at blog.Algorithmia.com for some examples.

Your platform is definitely interesting,and there's a lot of value locked in academic algorithms , but isn't the fact that most academics haven't offered algorithms via the cloud(it's not terribly complicated), hint to some other barriers ? And how do you plan to remove those barriers ?

I really love this approach. I am currently trying to do something similar for academics as part of my PhD project.

We have to understand that programming an algorithm actually is not always "easy". Imagine a new method presented in an academic paper that you would actually like to use. Re-Coding the whole thing that it works exactly the way presented sometimes is not possible (as not all implementation details are explained in the paper).

But imaging having a place (like Algorithmia) where researchers can upload there implementation (without giving access to their source code) and for others to access this algorithm with a very easy to use API would be a huge step forward in academics.

Why aren't academics just hosting in the cloud ?

Why aren't they just publishing the source?

many different reasons:

- no time

- no incentives (publish and perish)

- no knowledge

- usage of very different programming languages

- no platform to do this

Here's an initiative by Jimmy Lin to standardize implementations of IR used in Academia. Reproducibility of results is now made possible: https://github.com/lintool/IR-Reproducibility

Sounds more like $5 logos for programmers.

The technology behind it sounds really interesting - expecting programmers to chase bounties and work for nothing up-front with only the possibility of getting paid - not so much.

An algorithm defines a startup. Don't sell yourself short.

Creating accurate efficient algorithms is an art in that not anyone can do. However, the majority of problems the average engineer will run into has already been solved and are accessible with a little searching on google or stack overflow. Sites like this kinda make me sad.

It's called a 'SaaS marketplace' or 'API marketplace'

These people have come up here before, and I remain incapable of understanding who exactly is (a) going to be interested in using this and (b) is incapable of finding or implementing their own solution.

I though Algorithmia launched years ago?

It soft-launched in 2014, I believe. I registered on their website asking for an account sometime in 2014, and got an invite in September 2014 to what the email described as a "private beta". The general public launch where anyone can just click and create an account was earlier this year.

Mashape anyone?

Reply on news.ycombinator.com