HackerTrans
TopNewTrendsCommentsPastAskShowJobs

akisej

no profile record

Submissions

Databricks vs. Snowflake: The Battle to Power Enterprise AI

ashugarg.substack.com
2 points·by akisej·3 jaar geleden·0 comments

Florence Nightingale, datajournalist: information has always been beautiful

theguardian.com
2 points·by akisej·3 jaar geleden·0 comments

Data Distribution Shifts and Monitoring

huyenchip.com
1 points·by akisej·3 jaar geleden·0 comments

Show HN: AI SQL Copilot LogicLoop – AI to Generate, Optimize and Debug SQL

logicloop.com
71 points·by akisej·3 jaar geleden·45 comments

How to add Generative AI to your SaaS product

logicloop.com
2 points·by akisej·3 jaar geleden·2 comments

comments

akisej
·3 jaar geleden·discuss
Let me ask you something I always ask my reports: if you were your manager, what would you do?
akisej
·3 jaar geleden·discuss
Pretty interface, although I remain unconvinced of how I'd actually use it. If I'm just prototyping for myself, LLM providers offer a decent history, and I rarely need to share notebook-style explorations of LLMs with my team. For production use cases at logicloop.com/ai we just add our prompts into code.

What's the use case you're envisioning people using AI notebooks for?
akisej
·3 jaar geleden·discuss
Appreciate the effort you put into this. In terms of user flow, I typically need one or two of these use cases at a time. In that case, I just type my conversion into a search engine like Google, and often use their default box.

Can you share what types of use cases you've seen people use KodyTools for?
akisej
·3 jaar geleden·discuss
Yup, that's part of it but I mean it bidirectionally - users can accidentally leak data to models too, which is concerning to SecOps teams without a way to monitor / auto-redact.
akisej
·3 jaar geleden·discuss
These common issues tend to prevent LLMs from being used in the wild: * Data Leakage * Hallucination * Prompt Injection * Toxicity

So yes it does include prompt injection, but is a bit broader. Data Leakage is one that several customers have called out, aka accidentally leakage PII to underlying models when asking them questions about your data.

I'm evaluating tools like Private AI, Arthur AI etc. but they're all fairly nascent.
akisej
·3 jaar geleden·discuss
Great starting point! These diagrams notably miss a LLM firewall layer, which is critical in practice to safe LLM adoption. Source: We work with thousands of users for logicloop.com/ai
akisej
·3 jaar geleden·discuss
How about Amplitude or Heap? For an open source alternative you could consider PostHog.
akisej
·3 jaar geleden·discuss
Very cool, but curious if you see people actually directly interacting with LLMs vs in a script as part of a larger application? I see myself needing debugging, visualizing output etc. so much that an IDE makes more sense to me as an interface, so want to learn about cases where that doesn't.
akisej
·3 jaar geleden·discuss
Aw man, sorry to hear this about your friend. Inanimate objects are directly subject to the laws of physics, but living beings that have intention and will are able to circumvent those. For example, I can jump despite gravity existing. Yes, "in the long run, we're all dead", but applying laws of entropy as a reason to not live seems indicative of a lack of will, rather than a natural law every being must follow.
akisej
·3 jaar geleden·discuss
As long it's clear that this is fiction, how would something like this be more damaging than a series like The Man in the High Castle, or other sci-fi that imagines an alternate universe? I think it's a nifty technique that allows us to viscerally imagine and live out our parallel universes.
akisej
·3 jaar geleden·discuss
Yeah, in general the more data you're able to use (assuming the context window supports it), the better results tend to be. We arrived at the data schema being a good enough compromise at which the benefits outweigh the risks for several use cases. Besides, some data stores that are generated by third-parties actually have common schemas (think Sendgrid / Hubspot activity data), so you're not risking much but potentially gaining a lot of sales ops productivity.
akisej
·3 jaar geleden·discuss
This seems overall well-written and well-explained, but curious for that piece on fine-tuning. This article only recommends it as a last resort. That makes sense for a casual user, but if you're a company seriously using LLMs to provide services for your customers, wouldn't the cost of training data be offset by the potential gains you have and the edge cases you might automatically cover by fine-tuning instead of trying to whack-a-mole predict every single way the prompt can fail?
akisej
·3 jaar geleden·discuss
Thanks for sharing how you keep the discussion quality high here. We responded to some thought-provoking questions on this thread about differentiation vs ChatGPT, other SQL Copilots, edge cases like more complicated queries etc. that we believe other HN users and makers in this space will benefit from. As I responded in @chatmasta's vouch (thank you!), we're data and software engineering experts, and happy to provide our original perspective on any other questions you have for us. Thanks for your consideration!
akisej
·3 jaar geleden·discuss
On simpler multi-table joins we've been able to product good results, and we've done a lot of prompt engineering to make sure it takes the schema very seriously so that prevents hallucinations too. We're always finding new edge cases and fixing those as we go.
akisej
·3 jaar geleden·discuss
Thanks for sharing your experience. By organizing and storing queries, we hope to be able to improve these suggestions as well behind the scenes.
akisej
·3 jaar geleden·discuss
Yup, our bet is that people are going to want better integration into data sources and actions resulting from their queries, and that a lot of business value will come from that.
akisej
·3 jaar geleden·discuss
Yes, we can actually connect to third-party apps via APIs https://docs.logicloop.com/data-sources/supported-data-sourc...
akisej
·3 jaar geleden·discuss
That's a great idea for future improvements, thank you for trying it out!
akisej
·3 jaar geleden·discuss
Appreciate the vouch, @chatmasta! I actually did write those comments myself, and think that the 3-point explanation makes for a clear explanation. My cofounder and I both started our careers as software engineers, and I've been in the data space for over a decade.

A quick Google search for LogicLoop will also show you that we're both featured on Forbes 30 Under 30, and are funded by Tier 1 Silicon Valley VCs. https://www.forbes.com/profile/logicloop/?sh=67ca2b047a89
akisej
·3 jaar geleden·discuss
We use a combination of APIs from existing LLM providers, and do some serious prompt engineering to get the best from them. We're starting to train models on more SQL-specific prompts now.