Unsupervised Learning

On this week’s Unsupervised Learning, Pat and I sat down with CTO and Co-Founder of Sourcegraph, Beyang Liu. Sourcegraph is a leader in the AI coding space, and recently launched AI coding assistant, Cody. Beyang shared with us his view on the current landscape of AI coding and the future of coding and software development. He also shared how Sourcegraph has tried to make RAG better, and their model eval approaches.

(0:00) intro
(0:47) advice for young coders
(3:34) AI products at Sourcegraph
(6:17) the current state of AI coding
(12:33) what happens when a new GPT model comes out?
(20:16) what types of developers benefit from these AI tools?
(30:45) how important is inference cost?
(35:31) how does Sourcegraph structure AI teams?
(41:27) what metrics does Sourcegraph use to evaluate their products?
(50:02) customizing RAG
(56:55) getting ahead of the agentic future
(1:05:05) will there be more or less engineers in the future?
(1:13:50) over-hyped/under-hyped
(1:16:56) surprises during the Sourcegraph journey
(1:18:26) cognition buzz and Devin
(1:26:48) Jacob and Pat debrief

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

Pinecone has raised over $130 million and was most recently valued at $750 million. On this week’s Unsupervised Learning, we sat down with CEO and Founder of Pinecone, Edo Liberty. Pinecone is arguably one of the most important elements in today's modern datastack. Edo shared with us the most common use cases of Pinecone, the evolving landscape of vector databases, challenges in building vector databases, the "painful" launch of serverless model, and what people get wrong the most about Pinecone.

(0:00) intro

(0:33) what was it like when ChatGBT came out?

(6:29) Edo’s favorite applications built on Pinecone

(10:34) will we see more image and video applications in 2024?

(14:58) best ways to deal with hallucinations

(18:12) the evolving landscape of vector databases

(20:27) if Edo had to build a product, what would his stack look like?

(31:45) helping clients versus letting them figure things out

(36:38) moving to a serverless model

(40:33) what areas of AI should new startups target?

(45:18) Amazon SageMaker

(50:38) over-hyped/under-hyped

(51:30) biggest surprises while building Pinecone

(56:13) Jacob and Pat debrief

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

Mistral AI is often seen as the startup challenging OpenAI and incumbents developing LLMs. On this week’s episode of Unsupervised Learning, we sat down with CEO and Co-Founder at Mistral AI, Arthur Mensch. Arthur shared with us his view on why open-source will prevail, how Mistral gets LLMs into the hands of enterprises, the build vs. partnership decisions, the competitive landscape and future of LLMs, and how he’d regulate AI safety. 

(0:00) intro

(0:46) origins of the name “Mistral”

(2:20) logo origins

(3:06) closed-source vs open-source models

(6:31) “training models is what we do best”

(7:50) Mistral’s partnership strategy

(10:12) the next frontiers for LLMs

(11:47) Meta’s GPU announcement

(13:03) when will Mistral catch up to ChatGPT?

(16:00) NVIDIA chips

(16:55) AI regulation and EU AI Act

(20:07) who should handle AI safety?

(20:51) policy changes that Arthur would make

(22:52) foundation models around the world

(25:50) starting Mistral

(26:54) releasing Le Chat

(30:19) over-hyped/under-hyped

(30:32) surprises while building Mistral

(31:55) AI startups Arthur is excited about

(32:19) what application would Arthur build

(33:46) Jacob and Jordan debrief

With your co-hosts:  

@jacobeffron  

- Partner at Redpoint, Former PM Flatiron Health  

@patrickachase  

- Partner at Redpoint, Former ML Engineer LinkedIn  

@ericabrescia  

- Former COO Github, Founder Bitnami (acq’d by VMWare)  

@jordan_segall  

- Partner at Redpoint

Superhuman recently launched AI-powered Summarize and Instant Reply features, and has since processed 4 billion emails. On this week’s episode of Unsupervised Learning, we sat down with CEO and Founder at Superhuman, Rahul Vohra. Rahul shared with us what email will look like in the future, the internal product design decisions in building Summarize and Instant Reply, why he’s bullish on the agentic future, and why and how startups should go after incumbents.

(0:00) intro 

(1:20) why email will never die

(10:46) how ChatGBT changed Superhuman 

(17:01) making design decisions 

(24:34) how Superhuman personalizes email voices 

(28:35) choosing which models to use 

(31:00) how does cost play into decision-making 

(34:10) teaching users how to use AI 

(46:27) competing with incumbents 

(56:57) how work has evolved 

(59:02) how Rahul would redesign Slack / Slack Agent 

(1:04:54) Jacob and Pat debrief

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

There’s an ongoing debate about where the most value will accrue in AI between incumbents and startups. Of the incumbents, few have shipped product faster than SalesforceAI. Today on Unsupervised Learning we had on Clara Shih, CEO of SalesforceAI and one of Time Magazine’s 100 Most Influential People in AI. 

(0:00) intro

(0:50) work practices that will become irrelevant

(1:37) revolutionizing reply recommendations and case summaries

(4:57) newest Salesforce products

(5:53) structuring teams

(7:22) engineering trust into AI products

(11:58) combining in-house models with ChatGBT

(13:33) Gucci’s AI adoption

(16:01) how does Salesforce choose who to share their data with?

(20:29) AI costs

(26:29 creating unique voices for brands

(27:45) AI incumbents vs. startups

(29:54) what Clara would build if she had the time

(32:28) the future of Slack

(35:55) what percent of customer support questions can be answered by AI?

(38:37) over-hyped/under-hyped

(39:32) working with Mark Benioff

(40:46) Jacob and Pat debrief

(44:42) Slack is the perfect interface for generative AI

(46:10) Abridge investment

(48:15) Ideogram investment

With your co-hosts:  

@jacobeffron  

- Partner at Redpoint, Former PM Flatiron Health  

@patrickachase  

- Partner at Redpoint, Former ML Engineer LinkedIn  

@ericabrescia  

- Former COO Github, Founder Bitnami (acq’d by VMWare)  

@jordan_segall  

- Partner at Redpoint

Last week LangChain announced a $20M Series A led by Sequoia and released the paid version of LangSmith, which has already been used by 1K+ teams and driven 80K signups. On this week’s episode of Unsupervised Learning, we sat down with LangChain Co-Founder and CEO Harrison Chase to talk about the current state of LLM evaluation, observability, and the agent landscape.

(0:00) intro 

(1:07) applications of AI in the sports world 

(3:26) what does LangChain do? 

(7:51) building with LangSmith 

(10:00) best AI eval practices 

(16:51) to what extent is eval generalizable? 

(21:11) the current agent landscape 

(29:35) balancing present and future at LangChain 

(36:27) using LangServe to deploy LangChain applications 

(41:37) more complex chatbots are coming 

(45:51) current AI practices that will become obsolete 

(48:55) over-hyped/under-hyped 

(49:25) bigger surprise in building LangChain 

(51:50) how ubiquitous will open-source models be in the future? 

(52:43) most exciting AI startups 

(56:07) being an AI “celebrity” 

(58:09) Jacob and Jordan debrief 

With your co-hosts:  

@jacobeffron  

- Partner at Redpoint, Former PM Flatiron Health  

@patrickachase  

- Partner at Redpoint, Former ML Engineer LinkedIn  

@ericabrescia  

- Former COO Github, Founder Bitnami (acq’d by VMWare)  

@jordan_segall  

- Partner at Redpoint

Oscar Health is a $4B public healthcare company, providing healthcare insurance to nearly 1 million members. Oscar is at the forefront of AI adoption, continuously developing new AI use cases in healthcare. On this week’s episode of Unsupervised Learning, we sat down with Oscar Health Co-Founder, former CEO, and now President of Technology Mario Schlosser to talk about where AI will have the biggest impact in healthcare, top AI use cases at Oscar today, AI adoption challenges Oscar is facing, and limitations of GPT-4 in healthcare. Mario also shared his takes on open-source vs. off-the-shelf vs. healthcare-specific LLMs, and why can't we have robot doctors today.

(0:00) intro

(1:26) how will AI change healthcare in the next decade

(9:29) how Oscar uses AI

(19:00) how to build around healthcare requirements

(26:06) when would GPT-4 fail "miserably" and fundamental limitations of LLMs

(36:48) we shouldn’t piss off our smartest robots

(38:35) sharing AI knowledge between companies

(42:10) developing healthcare-specific models

(44:55) hackathons and karaoke nights at Oscar

(49:27) the need for a safety layer in LLMs

(51:53) best commercial opportunities in healthcare

(55:39) will their be AI doctors this decade?

(59:38) over-hyped/under-hyped

With your co-hosts:  

@jacobeffron  

- Partner at Redpoint, Former PM Flatiron Health  

@patrickachase  

- Partner at Redpoint, Former ML Engineer LinkedIn  

@ericabrescia  

- Former COO Github, Founder Bitnami (acq’d by VMWare)  

@jordan_segall  

- Partner at Redpoint

Replit raised nearly $100M at $1.2B valuation last April and powers over 20M developers. On this week’s episode of Unsupervised Learning, we sat down with Replit Founder and CEO Amjad Masad to talk about the future of software development, how Replit is empowering young users, how Replit developed its own models, and the data advantage Replit has. Amjad also shared his takes on why he’s bullish on agents, where the value in AI will most likely accrue and why open-source models might not be truly open today.

(0:00) intro

(0:45) advice for new coders

(6:20) how Replit uses AI

(10:36) AI’s coding capabilities

(15:49) what makes the best data

(20:52) educating new Replit AI users

(23:46) structuring AI teams

(27:02) building an in-house model

(36:54) “the world is gonna get way weirder”

(38:10) Kim and Taylor teaching calculus

(44:19) usage based pricing is going to get more prevalent

(51:05) will Microsoft win it all?

(55:00) Llama and vibe-checking AI models

(57:35) chatbots are overhyped

(58:18) latency matters

(59:50) why Sam Altman is the GOAT

(1:01:16) over 10 years we’ll see companies really shrink in size

(1:03:51) Jacob and Pat debrief

(1:05:27) training coding models on random data 

(1:06:36) Amjad’s take on agents

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

Intercom is one of the earliest adopters of AI - its AI product Fin has generated over two million answers and been used by thousands of users since it was launched last March. On this week’s episode of Unsupervised Learning, we sat down with Intercom Co-Founder and Chief Strategy Officer Des Traynor to talk about how AI is incorporated into Intercom, structuring its AI team, using RAG vs. fine-tuning techniques, where we are in the AI adoption curve, and his advice for startups building on top of AI. 

(0:00) intro 

(0:31) Intercom reaction to ChatGBT 

(3:18) how AI is incorporated into Intercom products 

(6:33) guardrails preventing hallucinations 9:49 exploration versus optimizing cost 

(19:00) structuring AI teams 

(31:17) fine-tuning for customers vs RAG 

(37:03) solving the 'actions' problem 

(38:38) lessons learned from transitioning into AI 

(44:28) over-hyped/under-hyped 

(45:53) companies that have implemented AI well/poorly 

(48:27) Jacob and Jordan debrief

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

OpenAI's inaugural DevDay sparked excitement in the AI community, with several product releases and ChatGPT hitting the milestone of reaching 100M weekly active users. On this week’s episode of Unsupervised Learning, we sat down with the Head of Developer Relations at OpenAI, Logan Kilpatrick. Logan shared with us how OpenAI prioritizes product builds internally, the interesting use cases he's seen for several OpenAI products, where OpenAI is headed, and what the Gemini release means for the ecosystem.

(0:00) intro

(0:33) how Logan uses ChatGBT

(1:36) underrated OpenAI products

(6:08) when is using GPT-4 necessary?

(7:22) custom GPT models

(9:05) are we at peak need for custom models in 2024?

(11:45) how does OpenAI prioritize products

(13:31) OpenAI’s text-to-speech model

(14:31) benefits of using open-source models

(21:00) what kind of company would Logan start if he left OpenAI?

(23:40) Google Gemini

(24:41) assistants API

(30:00) the need for a text-first AI-assistant experience

(35:18) putting limitations on agents

(42:18) the future of DALL-E and art generation

(48:00) over-hyped/under-hyped

(48:30) rare disappointments for OpenAI

(49:25) surprise successes for OpenAI

(50:03) how has OpenAI’s team developed?

(58:22) debrief with Pat

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

Perplexity is a next-gen search tool going after Google, with 1M Android app installs and 1M iOS installs within only 8 months of product launch. On this week’s episode of Unsupervised Learning, we sat down with the CEO and Co-Founder of Perplexity AI, Aravind Srinivas. Aravind shared with us the behind-the-scenes stories of how Perplexity AI was born (37:19), how he thinks about Perplexity being viewed as a "wrapper" (24:01), where search will be in 10 years (23:08), and where Perplexity is headed.

(0:00) intro

(0:48) the simplicity of Perplexity

(5:16) how Perplexity allocates resources

(7:09) don’t waste your time building your own models

(11:39) being a “wrapper” for OpenAI

(14:37) the future of Quora and Wikipedia

(19:38) what does it take to compete with Google

(23:08) what does search look like in 10 years

(24:01) showing users that Perplexity is more than a “wrapper”

(27:28) RAG solutions and solving hallucinations

(30:13) guiding users’ questions

(32:44) discover tab

(35:48) attracting new users vs. pleasing “power users”

(37:19) how Perplexity landed on search

(53:01) should AI be regulated?

(54:56) what other company would Aravind work at

(56:56) OpenAI

(58:42) Jacob and Pat debrief

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

In light of one of the biggest news stories in AI, we’ve put together a special episode to discuss the ramifications of Sam Altmans’ firing from OpenAI. Regardless of what happens between now and when you’re listening to this, the implications of the events that happened over the past few days are certainly worth unpacking. 

About our guests: 

• Alex Konrad, a journalist at Forbes covering Venture Capital and Tech 
• Jason Warner, former CTO at GitHub, partner at Redpoint, and now Founder of AGI start-up Poolside (https://www.poolside.ai/)

We go over what the implications are for AI Startups, who came away as the biggest winners, what’s going on behind the scenes and more. Hopefully you enjoy this episode. 

With your co-hosts: 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Linus Lee is an AI engineer at Notion, one of the earliest and most effective adopters of AI. In the episode, Linus shares how Notion developed its AI products, including Writer, Autofill, and Q&A, which just launched on Tuesday. It was fascinating to learn how Notion structures its AI team and dogfoods its development process. Linus also explores the hardest to anticipate when going to market with new AI features, and how Notion thinks of its LLM partnerships. Overall, a wide-ranging conversation about the behind-the-scenes stories of one of the most widely used AI tools today.

(0:00) intro

(0:37) T-Swift

(2:07) Notion AI

(9:08) approach to staffing

(16:51) educating users and user behavior

(22:32) challenges in developing Notion Q&A

(30:42) working with Anthropic and Open AI

(35:50) avoiding hallucinations

(36:23) switching AI models

(39:32) iterating on interfaces

(42:03) over-hyped/under-hyped

(48:03) Midjourney

(51:07) Pat and Jacob debrief

With your co-hosts: 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Jacob and Pat sit down with Erik Bernhardsson, the founder of Modal Labs, a data infrastructure company providing GPU compute to data teams. On this episode we discussed Erik’s thoughts on the AI chip market, the most popular GenAI use cases on Modal, and even Oracle Cloud’s resurgence in the AI start-up market.

0:00 intro

0:45 motivation for founding Modal

6:35 advantages that Modal gives developers

9:21 early applications built with Modal

11:58 challenges for AI developers

16:31 GPU access today

20:09 Vector DB companies

24:55 why is cloud adoption so slow?

31:30 Oracle Cloud

39:22 AI music generation

42:05 over-hyped/under-hyped

43:26 what Erik wishes he knew when starting Modal

45:53 episode debrief

With your co-hosts: 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Special Crossover Episode: We're excited to share this conversation from "The Logan Bartlett Show," another Redpoint podcast that focuses on untold stories from tech's inner circle. In the episode, Logan Bartlett interviews Dario Amodei (CEO, Anthropic) on the future of AI.

In the episode, Dario gives detailed predictions on the AI industry for 2024, 2025, and beyond. He discusses his days at OpenAI, leaving to start Anthropic, why he doesn’t like the term AGI, what AI developments he’s most excited about right now, and much more. Sharing here because we believe our AI enthusiasts will find this episode particularly enlightening. Enjoy!

(0:00) Intro 

(0:40) Joining OpenAI 

(13:51) Are scaling and AI safety intertwined? 

(19:51) Anthropic Early Days 

(23:24) Amazon's Investment in Anthropic 

(23:39) FTX investment in Anthropic 

(25:10) Anthropic's Business Today 

(30:11) Dario's Advice For Builders 

(33:14) Should we pause AI progress? 

(35:47) Future of AI 

(37:15) Dario's Biggest AI Safety Concerns 

(44:17) How Anthropic Deals With AI Bias 

(48:49) Anthropic's Responsible Scaling Policy 

(55:56) Testifying in front of Congress 

(58:45) Will AI destroy humanity? 

(59:20) GPT3 vs GPT4 

(1:01:46) The memification of a CEO 

(1:08:50) What are you most surprised by with AI? 

(1:16:23) Why don't you like the term AGI? 

(1:21:10) 2024 AI Predictions 

(1:33:05) Dario's opinion on open-source models 

(1:37:23) Probability of AI Catastrophe 

(1:40:12) Misuse of AI 

(1:44:04) Looking ahead: Dario's optimistic outlook on AI

With your co-hosts: 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Jacob and Pat sit down with Tome Co-Founder and CEO Keith Peiris to discuss Tome’s go-to market strategy, deciphering through the “AI tourists” to identify their ideal customer profile (ICP), and the different hardware cost considerations when reaching enterprise scale.

0:00 intro

1:35 founding Tome

5:05 designing Tome

10:30 how users want to interact with AI

12:26 teaching users how to use Tome

20:13 partnering with model providers vs. building your own

28:43 model evaluation

31:23 building an enterprise product

37:48 thinking about pricing

43:43 Keith’s favorite Tome projects

44:26 over-hyped/under-hyped

With your co-hosts: 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Jacob and Pat sit down with LlamaIndex CEO Jerry Liu to discuss his motivations for building LlamaIndex, thoughts on building enterprise-ready LLM applications and agents, and when fine-tuning makes sense.

0:00 intro

1:02 the evolution of LlamaIndex

3:48 apps being built with LlamaIndex

6:39 making agents more effective

12:58 retrieval augmented generation

16:49 what’s the right level of abstraction for LlamaIndex?

19:42 balancing reasoning and knowledge

30:46 storage for embeddings

36:03 underutilized features of LlamaIndex

40:38 over-hyped/under-hyped

With your co-hosts:

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

@jordan_segall

- Partner at Redpoint

Jordan and Erica sit down with Andriy Mulyar, Founder & CTO of Nomic AI, and discuss his motivation for creating GPT4ALL, the importance of data-centric AI, the use of LLMs in video games, and which technology companies are well positioned to “win” in the GenAI market long term.

0:00 intro

0:59 getting into AI and meeting Brandon

2:27 starting Nomic

7:43 how people are using Atlas

10:31 hallucinations in LLMs

13:05 gpt4all

17:25 building LLMs into video games

26:31 where does Nomic go from here?

37:39 Apple’s role in the LLM space

38:57 Andriy’s thoughts on AGI

40:41 over-hyped/under-hyped

42:13 AI regulation going forward

With your co-hosts: 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@jordan_segall 

- Partner at Redpoint

Patrick and Jacob sit down with Alexandru Costin, the VP of Generative AI and Sensei at Adobe, and discuss how Adobe’s early projects with generative AI in 2019 helped them move quickly upon the release of LLMs and diffusion models. Before leading Adobe’s generative AI efforts he founded InterAKT, a web development company, and led Adobe Romania for 10 years.

00:00 intro

01:44 Adobe Romania and background

02:26 AI projects at Adobe

10:03 incorporating AI into existing products

16:34 educating Adobe’s user base

25:13 avoiding copyright issues with Ai-generated content

32:12 using customer feedback

40:30 what is the right way to structure an AI team?

45:11 pricing AI products

48:00 the future of Adobe

With your co-hosts: 

@jasoncwarner 

- Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Jacob sits down with Alex to discuss how Snorkel grew from an open-source project in a Stanford AI lab to a $1B company. Alex shares his thoughts on why data development is at the heart of AI development, why enterprises are slow to deploy LLM applications, and the importance of academia in the future of AI development. 

00:00 intro 

01:03 moving from academia to Snorkel 

05:08 the evolution of Snorkel 

18:33 improving pre-training 

21:37 avoiding hallucinations and other errors 

33:00 barriers to enterprises deploying AI 

36:59 the Snorkel footprint of the future 

39:37 the role of academia in AI development 

42:57 over-hyped/under-hyped 

44:50 how should AI regulation change going forward? 

With your co-hosts: 

@jasoncwarner 

- Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

On today’s episode we talk with Jeff Huber, the CEO and Co-founder of Chroma. We talk about what sets Chroma apart from its competitors, new developments in AI technology, and advice for listeners who want to get started in AI.

0:00 intro 

1:02 starting chroma 

6:08 vector databases 

10:03 interesting use cases for vector databases 

13:14 what sets chroma apart? 

23:00 unresolved questions in LLMs 

32:45 multiple agents vs. one agent to rule them all 

34:50 chroma’s future 

38:00 embedding models 

43:00 over-hyped/under-hyped 

44:30 AI regulation 

48:42 is it too late to get into AI? 

With your co-hosts: 

@jasoncwarner 

- Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall 

- Partner at Redpoint

Jacob sits down with PathAI Ceo Dr. Andy Beck to discuss the state of AI adoption in diagnosis, why Path acquired their own lab, pathologists' jobs in the future and nailing GTM to reach a ~$1B valuation. 

00:00 intro 

01:01 pathology and AI 

13:30 nailing go-to-market strategy 

19:47 how pathology labs can go digital 

25:36 regulatory frameworks and roadblocks 

33:05 do improvements in foundation models impact PathAI? 

36:26 standardizing diagnosis 

40:05 how will the job of a pathologist change going forward? 

42:35 NASH 

47:30 working in Daphne Koller’s lab

With your co-hosts: 

@jasoncwarner - Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia - Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase - Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron - Partner at Redpoint, Former PM Flatiron Health 

@jordan_segall - Partner at Redpoint

Jacob and Jordan sit down with EleutherAI's Aran Komatsuzaki to discuss the future of open-source models, thought cloning, his work on GPT-J and more.

0:00 intro

01:06 Aran’s background

02:58 starting work on gpt-j

05:49 gathering data for Lion and gpt-j

08:51 history of EleutherAI

11:16 open vs. closed-source models

19:06 how will open-source models be used going forward

21:33 thought cloning

25:51 building AI models that understand video

29:35 one model to rule them all

31:58 influence of academia in the LLM space

34:33 over-hyped/under-hyped

38:01 Aran’s thoughts on AGI

With your co-hosts: 

@jasoncwarner 

- Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia 

- Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase 

- Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron 

- Partner at Redpoint, Former PM Flatiron Health

@jordan_segall 

- Partner at Redpoint

Patrick and Jacob sit down with Tatsu Hashimoto, Professor of AI at Stanford, to discuss the incredible open source projects from his research group like Alpaca and AlpacaFarm, whether data, algorithms, fine-tuning or RLHF is most important for performance, if AI is liberal or conservative, and much more!

(0:00) - intro

(1:05) - journey to Stanford

(2:50) - origins of Alpaca

(6:08) - capabilities of the Alpaca model

(16:39) - the future of AI

(20:07) - AlpacaFarm

(21:37) - how to improve language models

(29:15) - do language models form opinions?

(32:15) - how to solve bias in ai

(34:18) - how does academia fit into the world of AI

(42:01) - over-hyped/under-hyped

(46:35) - questions Tatsu doesn’t have time for

With your co-hosts: 

@jasoncwarner - Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia - Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase - Partner at Redpoint, Former ML Engineer LinkedIn @jacobeffron - Partner at Redpoint, Former PM Flatiron Health

@jacobeffron - Partner at Redpoint, Former PM Flatiron Health

Jacob sits down with Insitro CEO Daphne Koller to discuss founding Coursera, where and how ML can drive the most impact in drug development, and if foundation models can transform core drug discovery work and edtech. 

(00:00) - intro 

(00:54) - Daphne’s journey 

(09:18) - AI and biology discovery 

(10:59) - insitro vs. traditional pharma 

(20:04) - phenotyping patients 

(26:01) - early mistakes 

(29:51) - the future of data 

(35:33) - partnering with larger pharma companies 

(38:17) - impact of LLMs on biopharma 

(44:04) - over-hyped/under-hyped 

With your co-hosts: 

@jasoncwarner - Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia - Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase - Partner at Redpoint, Former ML Engineer LinkedIn 

@jacobeffron - Partner at Redpoint, Former PM Flatiron Health

Jacob sits down with OpenAI VP of Product & Partnerships Peter Welinder and guest host Rob Toews (Radical VC Partner) to discuss OpenAI’s strategy, how they think about what they will/won’t build, the future of open source models and when we’ll reach AGI. 

(00:00) - intro

(00:42) - where is the value in AI?

(07:51) - how OpenAI prioritizes projects

(14:15) - open-source AI

(25:49) - gaps in AI

(29:36) - risks and downsides

(34:02) - when will we reach super-intellegence?

(40:40) - super-intellegence safety

(43:37) - how OpenAI uses ChatGPT

With your co-hosts: 

@jasoncwarner - Former CTO GitHub, VP Eng Heroku & Canonical 

@ericabrescia - Former COO Github, Founder Bitnami (acq’d by VMWare) 

@patrickachase - Partner at Redpoint, Former ML Engineer LinkedIn @jacobeffron - Partner at Redpoint, Former PM Flatiron Health

@jacobeffron - Partner at Redpoint, Former PM Flatiron Health

Jacob sits down with ex-DeepMind / White House operator Teddy Collins to discuss Sam Altman’s Congressional testimony, how AI can enable direct democracy, state-driven economic planning and better orgs, US-China competition and more.

(00:00) - intro

(01:15) - how Teddy came to the AI world

(06:26) - working in government

(12:27) - AI regulation and Sam Altman testimony

(19:19) - the end of humanity

(34:34) - nearer term issues with AI

(41:10) - how the government can use AI

(47:13) - lessons from DeepMind

(54:54) - over-hyped/under-hyped

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Patrick and Jacob sit down with Mike Conover, Staff Software Engineer at Databricks and Co-Creator of Databricks Dolly, the world’s first truly open instruction-tuned LLM, to discuss the magic behind Dolly, Alpaca and other instruction-tuned LLMs, the unreasonable effectiveness of fine-tuning, how they got all Databricks employees to help them curate the Dolly dataset (hint: google forms), and more.

(0:00) - Intro

(5:54) - The birth of Dolly

(12:03) - Data curation at Databricks

(15:34) - Advice for building LLMs

(24:10) - The future of instruction-tuning datasets

(30:43) - UI innovation

(38:16) - The future of machine learning infrastructure

(42:05) - How SkipFlag would be different with the tools we have today

(47:01) - What Mike has learned since Dolly

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Jacob and Erica sit down with Jasper CEO Dave Rogenmoser to discuss the future of writing with AI and what it means for marketers and the internet, Jasper post ChatGPT, hosting a massive Gen AI conference, going upmarket and more.

(00:56) - How Jasper works and how people are using it

(02:00) - Dave's journey leading up to Jasper

(04:30) - The moment Dave knew he was onto something with Jasper

(07:30) - Where Jasper works well for business applications

(10:55) - How the content of the internet might change going forward with the use of AI for content creation

(13:00) - What a writer's workflow in the future looks like with AI

(14:50) - How the introduction of ChatGPT impacted Jasper

(17:46) - Looking at the long term opportunity for differentiation for Jasper

(23:16) - Tuning a 'brand voice' through Jasper's AI

(26:15) - Jasper's role in bringing community together, ie with their recent Gen AI Conference

(30:00) - How Dave thinks about the future of work for his kids

(32:20) - The productivity boost clients see through using Jasper

(37:00) - Quick fire round

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Jacob and Erica sit down with You.com CEO and former Salesforce Chief Scientist Richard Socher to discuss building a new search engine and the future of search, his predictions on when open source models will catch up to GPT-4, AGI and more.

(1:29) - Richard's career moves and what motivated them

(6:44) - More about You.com and the ethos behind the company

(9:47) - How You.com integrates 3rd-party apps

(13:30) - Richard's thoughts on the future of AI

(23:19) - The role of academia in the world of AI going forward

(27:30) - Things in AI that Richard is excited about

(31:53) - Quickfire Round

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Jordan and Erica chat with Fixie.ai co-founder and former Harvard CS professor on the journey from researcher to Google eng leader and then to startup life prior to founding Fixie, a new platform for building LLM-based apps (a Redpoint portfolio company). We also talk about having Mark Zuckerberg in his CS class, ChatGPT Plugins and the AI ecosystem, and what the future might look like for our kids with AGI. You can find Matt on Twitter (@mdwelsh) and learn more about Fixie at https://www.fixie.ai/

(1:06) - Matt talks about his early days at Berklee and Harvard

(6:45) - Making the jump to the start-up world

(8:15) - Matt explains what Fixie does

(11:30) - How Matt thinks about use cases for Fixie

(14:20) - How work might change with AI's integration

(17:04) - The future of LLM's

(20:30) - Matt's take on the race to AGI 

(25:20) - How Matt thinks about the future of the world for his kids and how they might use AI

(29:24) - Quick fire round

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Jacob and Jason sit down with NEAR CEO and Transformer paper author Illia Polosukhin. They discuss Illia’s fascinating journey from Ukraine to Google and AI to crypto, the origin story behind the “Attention Is All You Need” paper and the overlap between AI and crypto. Illia also shared his thoughts on AGI and the problems that excite him most in AI right now. You can find Illia on Twitter (@ilblackdrago) and learn more about NEAR (@nearprotocol)

(00:39) - Welcoming Illia; how he became interested in AI, transitioning into Crypto and explaining NEAR 

(02:40)  - Walking through Illia's story more in depth

(07:46) - How the Transformer Paper came to be

(11:24) - Understanding the Transformer Papers' impact

(18:28) - The overlap of Crypto and AI and how Illia sees the future of how they develop together

(26:47) - Illia's views on AGI

(30:46) - Optimism vs pessimism of the future of machine learning as a tool

(41:32) - What problems Illia sees in AI right now

(45:03) - Rapid fire questions

(47:36) - Where to learn more about NEAR and Illia

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Patrick and Jacob sit down with Matei Zaharia, Co-Founder and CTO at Databricks and Professor at Stanford. They discuss how companies are training and serving models in production with Databricks, where LLMs fall short for search and how to improve them, the state of the art AI research at Stanford, and how the size and cost of models is likely to change with technological advances in the coming years.

(0:00) - Introduction

(2:04) - Founding story of Databricks

(6:03) - PhD classmates using early version of spark for Netflix competition

(6:55) - Building applications with MLFlow

(9:55) - LLMs and ChatGPT

(12:05) - Working with and fine-tuning foundation models

(13:00) - Prompt engineering here to stay or temporary?

(15:12) - Matei’s research at Stanford. The Demonstrate-Search-Predict framework (DSP)

(17:42) - How LLMs will be combined with classic information retrieval systems for world-class search

(19:38) - LLMs writing programs to orchestrate LLMs

(20:36) - Using LLMs in Databricks cloud product

(24:21) - Scaling LLM training and serving

(27:29) - How much will cost to train LLMs go down in coming years?

(29:22) - How many parameters is too many?

(31:14) - Open source vs closed source?

(35:19) - Stanford AI research - Snorkel, ColBERT, and More

(38:58) - Matei getting a $50 amazon gift card for weeks of work

(43:23) - Quick-fire round

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health

Jacob and Jason sit down with Hugging Face CEO Clem Delangue and discuss trends in who’s using Hugging Face, the future of closed source vs open source in machine learning, why Clem compares large closed foundation models to Formula One Cars, how enterprise AI teams will evolve and AI safety.

(0:00) - Introduction

(1:37) - Welcome Clem

(1:57) - Starting Hugging Face

(5:42) - Influence of ChatGPT

(15:47) - Use cases of large vs. small platforms

(18:44) -  Should large language models be open?

(30:20) - What’s next for Hugging Face?

(43:06) - Rapid fire

(47:02) - Learn more about Hugging Face

With your co-hosts:

@jasoncwarner

- Former CTO GitHub, VP Eng Heroku & Canonical

@ericabrescia

- Former COO Github, Founder Bitnami (acq’d by VMWare)

@patrickachase

- Partner at Redpoint, Former ML Engineer LinkedIn

@jacobeffron

- Partner at Redpoint, Former PM Flatiron Health