Eugene Yan Profile picture
Apr 29 3 tweets 1 min read Read on X
Tenets from Duolingo's push to be AI-first

• AI will be everywhere in our product
• Start with AI for every task
• Spend 10% of your time learning
• Share what you learn
• Avoid overbuilding
• Build and experiment carefully
• Technical excellence still matters Image
Shopify had a similar push earlier this month
I adopt similar tenets at work, and folks who're hungry to learn & apply have seen gains of 2-20x.

Similar trends in the past include:
• Proprietary (sas/spss)->oss/python
• Adopting spark
• Machine learning, then deep learning

AI-first is another shift; don't overlook it.

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Eugene Yan

Eugene Yan Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @eugeneyan

Oct 30, 2024
Evaluating LLM output is hard. It's the bottleneck to scaling AI products for many teams

A key mistake is defining eval criteria w/o actually LOOKING AT THE DATA. This leads to irrelevant / unrealistic criteria + wasted effort

Thus, I built AlignEval AlignEval.com
The key insight: In addition to aligning AI to human context (prompting, rag), we also need to calibrate human criteria to actual AI output.

By working backwards from the data, AlignEval helps you build better evals.

Screenshots & how it was built here: eugeneyan.com/writing/aligne…
AlignEval simplifies building LLM-evaluators:

• Upload a csv with columns for input, output, and optionally, label
• LOOK AT YOUR DATA and label it with pass/fail
• Define eval criteria, run LLM-evaluator, eval the evaluator
• Improve your LLM-evaluator with "Optimize Mode" Image
Read 6 tweets
Jun 18, 2023
Our paper club recently revisited some of the earlier language modeling papers. Here's a one-liner for each.

---

Attention: Query, Key, and Value are all you need*

*Also position embeddings, multiple heads, feed-forward layers, skip-connections, etc

arxiv.org/abs/1706.03762
GPT: Decoder is all you need.

(Also, pre-training + finetuning 💪)

cdn.openai.com/research-cover…
BERT: Encoder is all you need. Left-to-right language modeling is NOT all you need.

(Also, pre-training + finetuning 📈)

arxiv.org/abs/1810.04805
Read 8 tweets
May 8, 2023
Ran a simple benchmark (Mandelbrot sets) between Mojo & Python. The speedup is impressive, and it benefits from Python's libraries.

• Python: 1,184ms
• Mojo: 27ms 🤯
• Python (vectorized): 240ms
• Mojo (vectorized): 2ms ImageImageImageImage
Here's a GitHub gist if you want to try it yourself: gist.github.com/eugeneyan/1d2e…

(Couldn't download the notebook for some reason)
Also hear what Jeremy Howard has to say about Mojo

Read 5 tweets
May 7, 2023
An insider's view on China's scale and tech, the 996 work ethic, and Alibaba's acquisition of Lazada. corecursive.com/software-world…

Years later, I'm still boggled by the scale and how we had to use a completely different tech stack (spoiler alert: it's mostly Ali Java). Image
Yea, there were one-click deploys, rollbacks, A/B tests—you name it.

Also, there were SQL commands that were both powerful and scary (and borderline questionable 🙈). Any data analyst on the street became a median data scientist. Image
The work ethic was punishing. Burnout became more common. While most Asians could endure it, folks from cultures that emphasized work-life balance struggled. Image
Read 5 tweets
Apr 11, 2023
Over the past few weekends, I've experimented with using LLMs to build a simple assistant.

Here's a write-up of what I built, how I built them, and their potential. Also, some shortcomings with embedding retrieval, with solutions from search & recsys.

eugeneyan.com/writing/llm-ex…
Here's my first project dabbling with LLMs for the humble `summarize`

Moving on to using tools, specifically SQL and search

Read 5 tweets
Apr 3, 2023
This weekend, I had a blast building a personal board of advisors using BeautifulSoup, @LangChainAI , and @pinecone.

`/board` provides advice to questions on tech, leadership, and life. It also provides links to sources for further reading!
`/ask-ey` does something similar for my own site, eugeneyan.com. And because I'm more familiar with my own writing, I can better spot shortfalls such as not answering based on a source when expected, or when a source is irrelevant.
A high-level overview:
• Scrape content from board of advisors (requests, BeautifulSoup)
• Embed content aka sources (OpenAI text-embedding-ada-002)
• Embed queries & find similar sources (Pinecone)
• Provide sources as context for the LLM to synthesize a response (LangChain)
Read 15 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(