For anyone not familiar, Raif is a Rails engine for building LLM powered apps. Highlights include:
- adapters for various LLM providers
- high-level abstractions/models for working with LLM's (Raif::Task for single-shot tasks, Raif::Conversation for chat interfaces, and Raif::Agent for building agentic features)
- web admin for viewing/debugging LLM requests/responses
v1.1.0 highlights include:
- Support for images and files/PDF's in Raif::Task's
- Embedding generation
- OpenRouter, GPT-4.1, and Claude 4 support
- Stats section in the web admin
- Automatic retries for LLM requests that resulted in server errors
Another Cultivate Labs person here. You're absolutely right re: predicting stock prices. A prediction market isn't going to beat the market consistently. People looking for that use case end up disappointed.
Prediction markets are more valuable to companies for things where there isn't already a large, liquid source of information flow. E.g. all the on-the-ground workers know a project is going to be late, but no one speaks up due to fear/office politics. Management thinks all is well since no one said otherwise.
A PM is an effective way of taking a bunch of siloed information/knowledge (in employee's brains), aggregating it, and presenting it to decision makers.
There was an article in NPR a while back about the Good Judgment Project (which was run by Dr. Tetlock - mentioned in the article). In some cases, betting/prediction markets (& other similar tools) were beating intelligence analysts - http://www.npr.org/sections/parallels/2014/04/02/297839429/-...
I participated in GJP for a couple seasons. Honestly, I didn't do that well.
But I was approaching it very differently than I would if it was my job. Since the system that was set up gave high rewards to unpopular predictions, I just gambled on the few most unpopular that had at least some shot at reversing. It wasn't the smartest approach, but it was the most fun. If I was doing it for real, obviously I'd go a different way.
The people who did best, at least from what I saw, tended to ride waves of popularity on the more active questions, buying low and selling high.
What I did was like betting on a few biotech startups, what the best scorers did was like riding waves of the market leading stocks.
In the end, I'm not that sure it had much to do with actual prediction of events. Then again, neither did my approach. I guess I'm not sold on the version of prediction markets they were using.
Kinda depends on which platform you were on (GJP used several). Some were prediction markets while some were opinion pools, both of which are scored/rewarded differently.
Our company (Cultivate Labs) recently acquired Inkling Markets (a very early YC company that built prediction market software) and have been building a new version of the PM platform, which will hopefully address some of the risk/reward quirks.
I'm about half-way through the book. Very easy, enjoyable read. Left wondering why some of these Amazon reviews try to make it sound as if it's quite academically/technically rigorous.
We typically only put one app on a server, especially since Digital Ocean is so cheap. If things are tiny apps, we just stick them on a $5/mo 512mb droplet.
100% agree with you. FWIW, this is much more heavily tested on ubuntu 12.04. One of the reasons we submitted it was the hope that we'd get more people to test & contribute. Goal being that you don't necessarily have to pay for heroku & can still own your servers, but can set up a new server in just a few mins.
If you run into hiccups, feel free to hit me up directly - [email protected]. The plus side of this is that once you do get it figured out once, setting up other servers really is extremely quick.
SEEKING FREELANCER - Chicago, IL - remote possible
Federis Group
We're a small software consultancy looking for front end and full stack engineers. Rails and/or mobile experience a plus, but not strictly required. Our team is based in Chicago, but we're open to remote as well.
We're a small software consultancy looking for front end and full stack engineers. Rails and/or mobile experience a plus, but not strictly required. Our team is based in Chicago, but we're open to remote as well.
Good call. Admittedly, I put this together very quickly yesterday when we ran into the issue of wanting to share a link. Would be nice if you could somehow avoid rendering a whole page, but I can't think of anything off the top of my head that would allow you to do that.
Instead, you can redirect to a web page with an iOS Smart Banner at the top. If the app is installed, the user must tap the open button. If the app is not installed, the user can tap the install button and the app will then install and open to the same designated in-app URL.
An alternative hacky method to avoid the web page is to redirect the browser to a JavaScript URL. However, this results in a momentary error popup, though it's automatically dismissed. And the user is still navigated to the backup URL when Safari returns to the foreground in the future.
EDIT: After thinking about it, since timers are suspended when Safari enters the background, you can probably use setInterval and check how long it's been to decide whether the browser should navigate to the backup URL. The error popup remains a problem, however.
Unfortunately, not yet, but we definitely hope to eventually. We just recently finished up the initial version of the iOS app, which was our first priority. Don't have the manpower yet to take it more than one step at a time :)
Currently, everything is based on virtual currency, so it isn't really gambling. It's primarily meant to be about entertainment, not about gambling with real money.
For anyone not familiar, Raif is a Rails engine for building LLM powered apps. Highlights include: - adapters for various LLM providers - high-level abstractions/models for working with LLM's (Raif::Task for single-shot tasks, Raif::Conversation for chat interfaces, and Raif::Agent for building agentic features) - web admin for viewing/debugging LLM requests/responses
v1.1.0 highlights include: - Support for images and files/PDF's in Raif::Task's - Embedding generation - OpenRouter, GPT-4.1, and Claude 4 support - Stats section in the web admin - Automatic retries for LLM requests that resulted in server errors
Full changelog is here: https://github.com/CultivateLabs/raif/blob/main/CHANGELOG.md