pklee
pklee on Hacker News
Very well deserved !! Amazing contributions !!
This is pure gold !! Thank you so much eugene and gang for doing this. For those of them which I have encountered, I can 100 % agree with them. This is fantastic !! So many good insights.
This is pretty cool. I tried with dolly and then I tried with T5-base, both of it did not give me result. It broke for me. Has anyone tried it ?
This is fantastic. This is what we do on our platform. + Copy and keep what worked and what did not so you can fine tune :)
If this is using OpenAI which it seems is what it is using, It is only sending column headers / column names. Not the data. If you are concerned about column names, you could also mask it on the way out and back in. If you are looking for an end to end database connect and query, please reach out to me.
They are decently good, I could not find major differences for the cases I was trying. The key is to control the temperature. Make sure it is low, otherwise the randomness increases tremendously. Infact you can feed the same input from openAI into NEOX and it generates results. There are many NEOX open playgrounds that allow you the control the temperature etc.
GPT3 model generates a SQL. You can sqldf on top of your data.table. We will be demo'ing at one of the events shortly. BTW, you could do somewhat similar with other LLMs such as GPTJ and GPT NEOX if you have worked with them
Very well said !! data is code
2 chapters in. Very good and makes you think. Truly impressive way of teaching how data is pivotal. The introduction reminds me of Peter Norvig's talk. I also like the point below of data is code, it is very profound.
The concept of a data-mesh is more of a business concept as opposed to tech. IMHO the idea being proposed is that of a conceptual data-server (not to be confused with database server) much like a HTTP server / Mail Server where people can engage with data as a first class citizen and create "data" products. This is especially true as we move from HTML to somewhat HDML (Hyper data markup).
By making data as the product (abstracting all the gory details), you are fundamentally engaging with data through a UI or an API. As you expose these products they become accretive while fundamentally encapsulating the domain expertise within them.
Of course. Assume somewhat of a big services company. There are many customer related problems that can be solved using products. Assuming such a product can be created. What should be the structure of the team that handles such products ? What should be the equity structure etc. Does that help ?
Posted by
pklee on Thursday December 23, 2021 @11:08AM
from the ask-hn dept.
What is the best way to structure teams ? Create equity ? Establish as a business unit ?
That was an awesome video. Thank you for sharing !!
What do you mean by dynamic scheams ?
This looks great. Thanks for sharing. I hope at some point this becomes a gate for RMF process.
I love the phrase "Enthusiasm is not a skill". I think it is great. I am not for or against in terms of your idea. I just wanted to comment on the phrase :)
Posted by
pklee on Monday June 21, 2021 @08:56AM
from the gartner.com dept.
Posted by
pklee on Monday June 14, 2021 @10:34AM
from the self-post dept.
I was looking at how people are thinking between GUI and datascience and came across this thread - https://news.ycombinator.com/item?id=16845666. In the past I have read GUI is basically a reification of command-line. Do we still think this way. Just what do we need as GUI for datascience.
Sorry I had meant this for @curiousperson23. Thanks
Yeah I am not just talking about analytics.. even basic merging across data-sources and simple visualizations based on the blended data and the ability to go across environments.
Yes completely. What I am seeing more and more is people try to shovel everything into big-data, but it becomes a nuclear powered can-opener which I am not sure is needed. The UI for this oscillates between excel with a macro or here is SaaS. Absolutely nothing in between.
Posted by
pklee on Thursday May 27, 2021 @12:04PM
from the ask-hn dept.
A lot of innovation happens in the big data (data that needs distributed compute - spark etc.) like ability to blend data across sources, schemaless / schema on the fly, deploy analytical models to production etc. Is there a case for similar innovations in the small to medium data (working with ~10M dataset) blended across data sources, simple analytical models and such ? What percentage of...
I would say Scott Ambler. His book "building object applications that work" was amazing and it helped me understand how to design systems
Posted by
pklee on Wednesday September 23, 2020 @12:04PM
from the self-post dept.
One of the top magnet school in USA, TJHSS wants to remove merit based admissions based on aptitude test and wants to replace it instead with lottery system. https://www.fcps.edu/news/superintendent-presents-recommendations-improve-diversity-tjhsst-establishing-merit-lottery