5 ms·

Just a quick reminder; my MIT licensed open source project Resin [0] does pretty much of what WA does. It's free today and it will be free tomorrow. Free to use

Just a quick reminder; my MIT licensed open source project Resin [0] does pretty much of what WA does. It's free today and it will be free tomorrow. Free to use in production, to clone, fork, tear apart. Like WA - it's a linguistic engine - it has a query language - there is a cloud you can use [1] if you don't want to host it on-premise - it has high level API's for creating language models from JSON documents (or in other ways formatted) - low level API's for creating whatever 64-bit vector space you might want - tools that help you in the analysis phase - tools for creating and deploy HTTP endpoints for reading from/writing to your model Next week I'm freezing the API's for I will have reached v0.4, because by then, at my day job, we will have completed our first "AI powered" closed-sourced product based on Resin. For this project to have any bearing and to stay alive and well (and free forever) I need your help. I need a community of contributors. Chip in, it's for a good cause! Edit, and here are the links: [0] https://github.com/kreeben/resin https://github.com/kreeben/resin [1] https://didyougogo.com https://didyougogo.com (currently down, b/c maintenance)

deckar01 7y ago
> does pretty much of what WA does... it's a linguistic engine Wolfram Alpha is a symbolic computation engine. I'm sure it does some natural language parsing for queries on the website, but I doubt that is an important feature for most developers considering this engine for production applications.
- misterman0 7y ago
  >> Wolfram Alpha is a symbolic computation engine. That's a great piece of technology to have when you are trying to get machines to understand humans. The reason Resin is not a symbolic computation engine is because that model hasn't been designed yet. Luckily Resin is a vector space framework that you can use to nest spaces together or transform into new spaces. Building on Resins capabilities for creating models from any 64-bit embedding I'm not quite sure if there is anything Resin cannot model.
  - reikonomusha 7y ago
    Yeah, but, come on. Isn’t this oversimplifying? “My computer uses bits. That means anything you can make out of bits can run on my computer.” It’s true but essentially vacuous. The value is/includes these difficult-to-build technologies, so the mere fact they could be included isn’t so meaningful and doesn’t add much value.
    - misterman0 7y ago
      Sure, but I was not trying to bluff you. Instead I was trying to describe to you a crucial point that is missing from the project's documentation (where I mostly describe the NLP engine) and that is that Resin at its core is a tool for designing such complex models as symbolic computation. But I hear you. Resin is not Wolfram but more of a framework that you can use to produce something like Wolfram.
dallamaneni 7y ago
I think you are confusing Wolfram Language (WL) with Wolfram Alpha (WA). > my MIT licensed open source project Resin [0] does pretty much of what WA does Does it? Resin looks more like an NLP framework at a quick glance. WA is an actual answering engine with data included. I fail to see a comparison. Not to mention the website is down.
- misterman0 7y ago
  >> I fail to see a comparison. Resin is a vector space framework. Upon that base I also built a NLP engine. And a chat bot. A search engine. Also, a anonymous profiler. So, yeah, it's more than an NLP engine. But it's not exactly Wolfram. You're right about that.
turingbike 7y ago
Regardless of similarity to Wolfram products, this looks cool. I feel like I could be in the target audience, but the documentation is quite sparse. I feel like > Your data ends up being represented as nodes in a graph that can be traversed by comparing a query (which is also represented as a vector) to nodes from the graph and use their cosine angle as a guide as to what node to traverse to next and when to stop. Could be... at least a page, with some diagrams. Each node... is a vector? Has a property that is the vector?
- misterman0 7y ago
  Thanks! The documentation is very incomplete. I'll fix that ASAP.
zwaps 7y ago
Can it analyze dynamical systems, solve complicated integrals, reduce and solve equations, display dynamic plots of all these things, do social network analysis, statistics, machine learning and data wrangling all in one notebook, out of the box, with simple calls? As a reminder, nothing in the Wolfram universe is really unique, nor it is the best individual implementation. The selling point of any Wolfram product and its underlying engine is that it can do all these things out of the box, in one notebook or WA query. I am not trying to be an ass. It just really seems you misunderstand the "competition". There is a failure path for your project: It COULD do anything, but it actually CAN'T do much out of the box. This limits your target audience to people who want to implement WA themselves, but with a framework. Please do consider this is not what WA offers. At all.

deckar01 7y ago

> does pretty much of what WA does... it's a linguistic engine Wolfram Alpha is a symbolic computation engine. I'm sure it does some natural language parsing for queries on the website, but I doubt that is an important feature for most developers considering this engine for production applications.

misterman0 7y ago
>> Wolfram Alpha is a symbolic computation engine. That's a great piece of technology to have when you are trying to get machines to understand humans. The reason Resin is not a symbolic computation engine is because that model hasn't been designed yet. Luckily Resin is a vector space framework that you can use to nest spaces together or transform into new spaces. Building on Resins capabilities for creating models from any 64-bit embedding I'm not quite sure if there is anything Resin cannot model.
- reikonomusha 7y ago
  Yeah, but, come on. Isn’t this oversimplifying? “My computer uses bits. That means anything you can make out of bits can run on my computer.” It’s true but essentially vacuous. The value is/includes these difficult-to-build technologies, so the mere fact they could be included isn’t so meaningful and doesn’t add much value.
  - misterman0 7y ago
    Sure, but I was not trying to bluff you. Instead I was trying to describe to you a crucial point that is missing from the project's documentation (where I mostly describe the NLP engine) and that is that Resin at its core is a tool for designing such complex models as symbolic computation. But I hear you. Resin is not Wolfram but more of a framework that you can use to produce something like Wolfram.

dallamaneni 7y ago

I think you are confusing Wolfram Language (WL) with Wolfram Alpha (WA). > my MIT licensed open source project Resin [0] does pretty much of what WA does Does it? Resin looks more like an NLP framework at a quick glance. WA is an actual answering engine with data included. I fail to see a comparison. Not to mention the website is down.

misterman0 7y ago
>> I fail to see a comparison. Resin is a vector space framework. Upon that base I also built a NLP engine. And a chat bot. A search engine. Also, a anonymous profiler. So, yeah, it's more than an NLP engine. But it's not exactly Wolfram. You're right about that.

turingbike 7y ago

Regardless of similarity to Wolfram products, this looks cool. I feel like I could be in the target audience, but the documentation is quite sparse. I feel like > Your data ends up being represented as nodes in a graph that can be traversed by comparing a query (which is also represented as a vector) to nodes from the graph and use their cosine angle as a guide as to what node to traverse to next and when to stop. Could be... at least a page, with some diagrams. Each node... is a vector? Has a property that is the vector?

misterman0 7y ago
Thanks! The documentation is very incomplete. I'll fix that ASAP.

zwaps 7y ago

Can it analyze dynamical systems, solve complicated integrals, reduce and solve equations, display dynamic plots of all these things, do social network analysis, statistics, machine learning and data wrangling all in one notebook, out of the box, with simple calls? As a reminder, nothing in the Wolfram universe is really unique, nor it is the best individual implementation. The selling point of any Wolfram product and its underlying engine is that it can do all these things out of the box, in one notebook or WA query. I am not trying to be an ass. It just really seems you misunderstand the "competition". There is a failure path for your project: It COULD do anything, but it actually CAN'T do much out of the box. This limits your target audience to people who want to implement WA themselves, but with a framework. Please do consider this is not what WA offers. At all.