There is just no way this could ever be a problem. Surely everybody knows that if presented with some data they need to do a deep dive into the actual sources instead of blindly trusting a graph.
I seriously want to know what was going on in the mind of the person who made this.
I think we've (https://www.definite.app/) done a decent job here. Once the LLM generates results, you can inspect the query all the way back to the source (SQL).
'daily unique visitors to openai.com by month since 2022 to 2024'
It gave a graph with 1) time axis decreasing left to right, 2) visitor numbers which can't be real (near prefect linear trend) 3) points in the future, going out to Dec 31, 2024.
I had to press the “refine” button several times which turned the query into “Bitcoin price 1900-2300”
The result was a bar chart with five data points.
All of the five data points were labeled “2K”, and had differing heights.
The dates along the X axis were in a short span of time – just a few days – and were seemingly randomly ordered. Or like, sorted with even dates in descending order followed by odd dates in ascending order. Very.. peculiar.
2023-10-04
2023-10-02
2023-10-01
2023-10-03
2023-10-05
Also the link copy button is not working for me in Safari on iOS, and the url doesn’t change from the op link either, so I can’t share a link to the graph.
> After trying some queries, I found ChatGPT is good at finding public data, sometimes it may not be accurate, but overall, pretty good.
GPT-4o is no more reliable for this than GPT-4 or GPT-3.5 before it.
One of the many unintuitive things about LLMs is that they are terrible tools for looking up specific facts about the world - which is weird, because that's traditionally one of the things computers have been best at.
If you want to build a working version of a system like this the trick would be to provide it access to tools for looking up data in citable, reliable data sources - things like the CIA World Factbook.
Then display the charts along with a link to that underlying reliable data.
Compare 4-turbo to 4o
llm -m 4t "Tell me about these cli tools and link to them: - llm: - ttok - symbex - shot-scraper " Among the tools you've listed, "shot-scraper" is the only one that I could provide detailed information about based on widespread use and documentation available as of the last update. As for "llm," "ttok," and "symbex," there isn't readily available, widely recognized information on tools precisely matching those names, so it's possible they might be lesser-known tools, newly released, or possibly regional or specialized tools not widely documented in widely-used platforms. (Truncated)
llm -m 4t "Tell me about these cli tools and link to them: - llm: - ttok - symbex - shot-scraper " As of my last knowledge update in September 2023, here is the information on the CLI tools you are asking about:
1. *llm*: - I couldn't find a specific command line tool named "llm" that is widely recognized or used. It could be a lesser-known tool, a new tool, or potentially a typographical error. Please verify the tool's name and provide additional details if possible.
2. *ttok*: - As with "llm," there isn't any widely known CLI tool named "ttok." (Truncated)
llm -m 4o "Tell me about these cli tools and link to them: - llm: - ttok - symbex - shot-scraper " Sure! Here are brief descriptions of the CLI tools you mentioned along with links to their official repositories or webpages:
### 1. llm `llm` is a command-line interface (CLI) tool to interact with large language models like OpenAI's GPT-3. It allows you to query these models directly from the terminal, making it easy to integrate AI capabilities into your workflow or automate tasks.
- *Link*: [llm GitHub Repository](https://github.com/simonw/llm)
### 2. ttok `ttok` is a CLI tool for working with tokenizers. It's a tool under development aimed at helping users understand how tokenization processes text, which is essential for preparing data for machine learning models, especially in NLP applications.
- *Link*: [ttok GitHub Repository](https://github.com/simonw/ttok)
### 3. symbex `Symbex` is a CLI tool for performing symbolic execution on Python bytecode. It can analyze Python code to find potential bugs, vulnerabilities, or logical errors by exploring all possible execution paths.
- *Link*: [symbex GitHub Repository](https://github.com/simonw/symbex)
(Truncated)
llm -m 4o "Tell me about these cli tools and link to them: - llm: - ttok - symbex - shot-scraper " Sure, here is a brief overview of each CLI tool along with links to their respective repositories or documentation:
1. *llm (Language Model CLI)* - *Description*: The `llm` command-line tool leverages large language models, such as OpenAI's GPT-3, to make it easier to incorporate AI functionalities into your command-line tasks. You can use it to generate text, answer questions, and assist with coding or other language-based tasks directly from your terminal.
- **Link**: [llm on GitHub](https://github.com/simonw/llm)
- **Link**: [ttok on GitHub](https://github.com/simonw/ttok)
- **Link**: [symbex on GitHub](https://github.com/paulgb/symbex)Clearly, openai did a lot more web scraping for this model.
I'd urge you to take this site down as it will be net negative to the world.
- What is your core technical value add?
- Do you have the data sets in your own database and you are using OpenAI to query them?
- Looks like you have your own home built database?
- Are you using LLM Agents?
- I saw that you have Airtable integrations, are you able to do the same for any datasource including Airtable?
It does not use database for any "random search", but yes, columns.ai is a data analytics tool that allows you to connect supported live data sources like Google Spreadsheet, Airtable, Notion Database to create visual stories.
The analytics engine is home built (https://github.com/varchar-io/nebula) but it is not a database. And I don't use LLM agents, just build logic how to purify data returned by LLM, and fit them into an optimized visualization.
Hope I answered your question!
Anything is clickable and styleable on the canvas though.
4k each year but the bar chart goes up
Sounds about right
This site is a harbinger of the agitprop factories which are about to flood the 'net due to the general availability of LLMs.