Austin

Data Science Enthusiast
#DontPanic #Rule32

[charlie marketplace @charliemktplace]

Me (married w/ 2 kids & enjoying life) begging my single friends to stop "playing it cool" and "seeing how it goes" with their 15th situationship they actually have a crush on since Obama

[T-Pain @TPAIN]

Baby girl, what’s your name?

[Zach Wilson @Zachly]

I'm collaborating with Meri Nova (@intelligentle__) on creating a breaking-into-data handbook!

Inside this repository, you'll find links for courses, books, content creators, podcasts, and everything else you'd need to excel in the data domain!

This repository is 100% open source and open to suggested changes as well!

https://t.co/mlQLblOjwX

#dataengineering
#datascience

[harpreet @DataScienceHarp]

Alright, you've seen me post about the zoomcamp, and you've probably heard the name, but maybe you're not sure what 🦜🔗LangChain is and why it's worth learning...

In one sentence: 🦜🔗LangChain is a framework for creating language model-powered applications by chaining together language models and components.

⛓️ "Chain" refers to combining different modules and LLMs to build pipelines and workflows.

It's a sequence of calls to components that combine prompts, models, tools, and other components to create custom language model-powered applications.

What do I mean when I say "component"?

🧱 Think of these as building blocks.

Components (aka modules) provide interfaces and integrations for capabilities like interacting with models, data, constructing chains, etc.

They're modular, designed to be easily composed, and chained together to build custom applications.

They provide you - the developer - with reusable, composable building blocks that make it (relatively) easy to develop complex language model-powered applications.

Now, why should you care?

A language model by itself is awesome, but it becomes fully awesome when you can connect it to external data sources (ie, the internet or your private data) and let it interact, query, and otherwise work with that data.

This is driven by two core principles of the LangChain philosophy:

1) Data awareness. That means not limiting your language model to what it knows out of the box but expanding its horizons by connecting it to other data reservoirs.

2) Agentic. A language model should be able to interact with its environment actively. It's not just a passive tool; but a dynamic part of your app ecosystem.

Combine components with the core principles of LangChain and you have a framework that provides a simple way to jumpstart an application.

So there you have it - LangChain in a nutshell.

Tomorrow I'll provide an overview of the various components in LangChain, and over the next few days, we'll delve into each component in-depth and provide code where necessary.

[Gokul Rajaram @gokulr]

The power of high quality decision making

From an ex-colleague who’s now head of product at a growth stage company:

“wrote a SPADE doc yesterday 😉 tracked down your coda / first round post to explain what this framework was too.

tldr it went very well. Very well. Helped ensure we thought of all parties impacted (customers, financials, and internal teams), all were consulted and felt heard, the alternatives exercise enabled me to intelligently answer how other options had positive and negative tradeoffs, quantifying the impact of each option was hard but ultimately made what to decide very easy to understand and accept. Now we have to do the work to make it happen 😅”

[Thomas Frank @TomFrankly]

Had dinner with @JamesClear the other night. This insight came up, which I think is worth sharing:

Draw a box around all the things you're willing to do.

Within that box, do everything possible to maximize income and growth.

Don't do anything outside the box.

[Vin Vashishta @v_vashishta]

My unpopular opinion could save you from serious career setbacks. Everyone in a white collar role should be upskilling to Data Analyst capabilities. For executive and C-level leadership roles, Data Science capabilities will be an increasingly common requirement. Why? 1/7

[Jeffrey Li @jeffmli]

Great news! We just launched Profile Poet on @ProductHunt

Free, personalized, AI-generated dating profile.

We'd love all the support we can get. Try it out! https://t.co/NVrJ228bOA

[Dickie Bush @dickiebush]

ChatGPT is overhyped.

That's what I told myself after 2 weeks of trying (and failing) to use it well.

Turns out, I was just a poor prompt writer.

But after spending hundreds of hours tinkering, I've finally cracked it.

And now, it's my personal writing assistant.

Here's how:

[swyx @swyx]

ChatGPT casually dropped an APP STORE 🤯

It can now:
- browse the web (RIP Bing waitlist, cutoff)
- write and run Python (RIP replit?)
- access org info (RIP docsearch startups)
- add third party plugins from OpenTable, Wolfram, Instacart, Zapier, etc)
- developer SDK in preview

[GREG ISENBERG @gregisenberg]

The future of building startups:

- MVP speed (1x per month)
- AI-accelerated
- Superniche is the new niche
- Community 1st, software 2nd
- No-code 1st, some code 2nd
- 10x more automated
- Global teams, localized products
- 95% dominated by solopreneurs and microentrepreneurs (teams less than 12)
- Pop-up digital experiences (apps that only work on certain times)
- Needs the marketing holy-trinity to hit escape velocity: 1. product/market fit, 2. content/market fit and 3. community/market fit
- Team is half robots 🤖, half humans 👨‍🦰 (cc @youneedarobot)
- Accelerated by "boring marketing" (cc @boringmarketer)
- Multiple revenue streams
- Design matters. The bar is high
- Partnered w/ creators (creators are the distribution)
- Feels like a game (levels, status, badges, in-app currency, challenges, collectibles/items)
- Purpose-driven moonshots: societal impact matters
- Productized agencies to generate cashflow (ex design agency @DispatchDesign)
- Product studios become the norm
- 99% of MVPs won't need VC

[Radek Osmulski @radekosmulski]

Programming has changed forever 😳

Today I used GPT-4 for the first time and I can't imagine coding without it ever again.

Here is the project I worked on:

[Hasan Toor @hasantoxr]

GPT-4 will make your life super easy.

It is confirmed to be doing amazing things.

Here's everything you need to know about GPT-4:

[Aadit Sheth @aaditsh]

ChatGPT prompts that'll save you hours a day at work (ranked in order):

[Vin Vashishta @v_vashishta]

OpenAI dropped their prices this week for the 2nd time in a year. Why drop the price when they dominate this early market? Competition and adoption.
1/5

[Josh Tobin @josh_tobin_]

What I'm reading this week in production ML / LLMs 👇

[Daliana Liu @DalianaLiu]

You don't *really* know machine learning if all you do is model .fit() but don't do error analysis.

Here are 5 tips on how to do error analysis in 2023:

[Mickey (Miki) @BazeleyMikiko]

Thanks so much for the feature @hashnode for my #devretro2022 #devretro2022hashnode wrap-up post!

3 x Retreats (Austin, SLC, Miami)
2 x Conferences (SF, Toronto)
10 x panels & talks
6 x podcasts
16 x livestreams
5 x blog posts
Many new friends

#2022 memory thread 🧵

1/5

[Mickey (Miki) @BazeleyMikiko]

The last session of #AoDSHappyHours - although 😢 Fridays will no longer be chatting with other #datascientists #mlengineers #datastrategists, so many good friends made including @DataScienceHarp @KenJee_DS @v_vashishta @mellamomark_eth @smasis @bentaylordata etc

#2022retro

2/5

[Abacus.AI @abacusai]

Free machine learning crash course from @BerkeleyML.

Highly recommended for anyone looking to start!

https://t.co/ygHNkWxU9X