Too often, my Twitter experience now is - start reading a post with some kernel of insight - read some claudism in the post - stop reading immediately

It seems like people turn over their raw kernel of insight to an LLM to give it polish, likely in order to increase its appeal to the audience. But most of that appeal is lost amongst people that consume LLM written content regularly.

It's also likely that social media engagement does not filter for this. A majority of people are still not neck deep in LLM-talk, and cannot distinguish between LLM and human written content.

It'll probably take us another year or two to develop herd immunity against meat proxism.

Everything that's currently available will continue being free. We're introducing a tentatively $50/mo pricing tier for more data (insights from earning transcripts and leadership podcast interviews, bespoke KPIs from 10Qs, more macro and pricing data, insights from satellite imagery) - but are still figuring out what the right price point should be!

The problem

AI agents are great at data analysis. But they become ineffective if most of their context window is spent on gathering and cleaning data, instead of validating hypotheses.

Data in the wild is messy and rarely standardized. Definitions and measurements change over time. This problem is compounded by a fragmented data universe. Point solutions exist for getting just market data, just macro data, or just trade data. But unless you pay $30k/person/year for a Bloomberg terminal - you’re not getting that data inside your AI agent.

How we're fixing this

We source, clean, and standardize data from SEC filings, official economic series and macro-economic releases. We structure the data so it’s stupidly obvious to query for any agent. All data is stored in just 3 tables with identical schemas and a total of just 20 columns. We keep the metadata like units, methodology and coverage fresh.

We're launching today with plugins for Claude Code and Codex. The plugin is open source, so you can adapt it to other agents.

So far we have over 25M+ series including macro and detailed bilateral trade series from 8 countries, US equity prices and fundamentals, and much more.

We’re adding new data every day - with structured insights from satellite imagery, investor calls and company leadership interviews coming soon.

Our ask

If you go deep into company fundamentals or sectors, and use AI agents in your investment decisions, we’d love to chat. Please drop us feedback below or at founders@factiq.com!

Comments URL: https://news.ycombinator.com/item?id=48826422

Points: 7

# Comments: 2

Comments URL: https://news.ycombinator.com/item?id=46780730

Points: 1

# Comments: 1

Live Demo: https://factiq.com/

Datasets: BLS, EIA, BTS, Census EITS, and more

What FactIQ is - data analysis tool for 7M+ economic data series

- meant for folks who need quick access to reliable economic statistics and contextual analysis

- emphasis on data traceability and citations

For reference, here is a quick FactIQ analysis that compares how sources of electricity in the US and China changed over time: https://www.factiq.com/share/7c0bf9a6d8304c01849af763f39362e...

Why we built it

- we worked on Defog.ai for 3 years, selling fine-tuned AI data analysts to large enterprises based on our open-source text to SQL models

- saw non-trivial usage and adoption, but it became more of a consulting company instead of a scalable product, because we needed to spend a lot of time gathering context and understanding DB schemas for each customer

- decided it was a better bet to build something which was widely useful and where we could design and control the database schema to make it more AI friendly

- saw how broken the current state of economic data discovery was, and decided to fix it

Technical Details - we get large data dumps from government agencies, standardize them into a common internal schema, use LLMs to extract metadata and enrich the data, and create embeddings to make it searchable

- when users ask questions, we do an embeddings-based search (to discover relevant data series) along with an agentic analysis pipeline for actually answering questions

What's next - More comprehensive data about the US economy

- Adding data for China, India, and the EU

What we're looking for - Ask complex questions, and give us feedback on what we can improve

- If you work regularly with economic data, tell us what your workflow needs are

Comments URL: https://news.ycombinator.com/item?id=46120954

Points: 2

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=43969027

Points: 17

# Comments: 3

Comments URL: https://news.ycombinator.com/item?id=43483576

Points: 1

# Comments: 0

SQLCoder is a fine-tuned variant of StarCoder, supplemented with a lot of hand-curated of data and slightly novel fine-tuning techniques.

We are also open-sourcing our framework for evaluating whether LLM-generated SQL is correct. SQL is tricky to evaluate. Two very different SQL queries can both be "correct". For the question, "who are the 10 most recent users from Toronto", both of the following are correct in their own ways – so we had to build a new framework algorithm to evaluate query correctness.

Query 1: ```sql SELECT userid, username from users where city='Toronto' order by created_at DESC LIMIT 10; ```

Query 2: ```sql SELECT userid, firstname || ' ' || lastname from users where city='Toronto' order by created_at DESC LIMIT 10; ```

The model is small enough to run on a single A100 40GB with weights in 16 bit floats, or on a single high-end consumer GPU (like RTX 3090/4090) with 8bit quantization. We will also release a ggml-based quantized version soon, and you should soon be able to run it on most M1 or M2 Macbooks with 32GB of RAM.

The model weights have a CC BY-SA 4.0 license. You can use and modify the model for any purpose – including commercial use. However, if you modify the weights (for example, by fine-tuning), you must open-source your modified weights under the same license terms.

Our evaluation framework is at https://defog.ai/blog/open-sourcing-sqleval/ Interactive demo: https://defog.ai/sqlcoder-demo/

Would love for you to give it a spin, and let us know what you think!

Comments URL: https://news.ycombinator.com/item?id=37209539

Points: 5

# Comments: 1

Do you have plans to let users define "types" of data that can be redacted (like monetary terms in a contract, code embedded in documents etc)? Also, any plans on making this an API that other developers could build on top of?

Comments URL: https://news.ycombinator.com/item?id=36164681

Points: 1

# Comments: 0

I wonder how the devtooling around this will evolve. Seems like a matter of days until someone creates a GUI wrapper around this, and obviates the need to use programmer time for fine-tuning

Comments URL: https://news.ycombinator.com/item?id=33797112

Points: 1

# Comments: 0

[1] https://app.betafi.co/studies?create=true&template_id=p_wqzp...

[1] https://threadreaderapp.com/thread/1588016682060378112.html

[1] https://en.wikipedia.org/wiki/Portal:Current_events/November...

Comments URL: https://news.ycombinator.com/item?id=33402758

Points: 1

# Comments: 0

Linking to verified sources (research papers, official websites, verified social media accounts) when writing about these topics might make this easier. LLMs will then be able to understand if one if misrepresenting what was stated in the linked sources or not.