Lunar New Year - Fertility, CyberSecurity, DeepSeek | Erin Jerri
Technology
Lunar New Year Wrap-Up - All About AI (Women in AI Club Annual Pitch Celebration, AthenaDAO: AI x Fertility, New Club: Women in AI x CyberSecurity, Taro: The DeepSeek Debate)
Date Published
Reading Time
8 min read
tl;dr In this email - all about AI
Lunar New Year Greetings
In the spirit of all things Asian coming into the New Year, listen to the hot new track on Spotify, “Bakunawa” by my sis, scientist and emcee Ruby Ibarra. 🐉🎶
AI - 3 Things
AI for Women
LLM Agents and Reasoning - I temporarily shelved my posts on AI Safety and AI Productivity since another move was made by UC Berkeley’s Dawn Song, who just released a new course on Advanced LLM Agents and reasoning. Last term she had a great lecture about safety along with Anthropic’s Co-Founder, Ben Mann and Song will continue to expand on AI safety at the end of term. I will release posts in the future.
Join AthenaDAO- Learn more about how AI is used to improve fertility outcomes. I’m going to stake the ATH token. Anyone who wants to have children in the future (freeze their eggs and embryos, do IVF, get pregnant) and allies that seek to understand their options should take this course! Hands-down this is the most technical and scientific course I have taken that weaves in AI, web3, and DeSci - Decentralized Science.
Women in AI and Cybersecurity. I along with other senior engineers, startup founders, and investors got to ask questions in a rapid fire Q&A hot seat with the Chief Technology Officer (CTO) of Okta, Bhawna Singh. Apply to join The New Club, let them know I referred you.
I am both a descendant of the anti-Spanish colonial revolutionaries, the Katipunan (my great grandfather was a member Guillermo Pangilinan, on my dad’s side) and am 1/16 Chinese and 1/8 Spanish with both sides of my family from the Philippines (my mom is from Pampanga and Dad is from Manila). I have quite a few friends who are Chinese-Filipino and do actively celebrate Lunar New Year. Growing up in Fremont, I had plenty of Chinese and even Vietnamese friends who celebrate. In Filipino indigenous culture, we have tales of the bakunawa - the dragon serpent eating the moon.
Check out the recently released “Bakunawa” - Shout-out to my scientist musician sis, Ruby Ibarra, in collaboration with Ouida, Han-Han, the legendary June Millington and other members of Bolo Music Group and Pinay Voltron (shout-out to Anna).
I went to Women in AI’s Lunar New Year Celebration. We saw amazing pitches by four solid founders who are changing the game in all areas of AI, working on some of the world’s hardest problems. Women in AI founder, Claire Xie herself also showed a demo for her SaaS product company, Clink.ai (I love live demos).
What a great way to celebrate Lunar New Year with women pitching some amazing AI companies! 🧧 🐍
Outside of meeting doctors building their healthtech companies, I saw some very solid pitches by AI companies who are 3D printing houses (Anna Cheniuntai founder of Apis Cor), creating computer vision-enhanced apps to have healthier options for grocery checkout Anastasia Trofimova, founder of Your Beet) and are changing up the game in SaaS (Claire Xie founder of Clink.ai), structured search (Erika Bahr striking at it again with a great pitch for Daxe, and fighting deep fakes (the very fierce and inspiration founder of Breeze Liu, founder of Alecto AI .
It's only natural woman tackle these kind of hard problems I’ve seen are hard to solve, women are leading the way in AI and are absolutely killin’ it!
Shout-outs to the founder, Claire Xie (also a startup founder!) for her outstanding work at Women in AI Club - literally the one organization I have seen truly highlighting the very best women founders and investors in AI.
AI for Women
AI for Fertility - AthenaDAO - web3 Fertility MOOC
Shout-outs to AthenaDAO for finally releasing their fertility MOOC. I was so excited to finally take this course, after taking USC Professor/Dr. Cleopatra’s Fertility MOOC last year (pricey supplements with). Professor Cleopatra Abdu is the first African American professor at USC in the School of Gerontology and Psychology, who has been a trailblazer in this space. I’m privileged to have participated in her MOOC as well as the Prolon fast. This was foundational for me. Building upon my knowledge there and webinars by Spring Fertility in Silicon Valley, where I sought my fertility consultation, I was able to get a base amount of knowledge, but there were still gaps in my knowledge. Also, I was curious as to how other modern technological developments (AI and web3) apply to fertility given a lot of the growth of the Decentralized Science (DeSci) movement in recent years, with even Ethereum founder, Vitalik Buterin talking about his support for synthetic wombs I’m all for this btw in case I can’t have babies for whatever reason).
Last week’s lecture by Chiranjeev Bhatia, a highly skilled embryologist at IVF Londondescribed how AI is used in a variety of areas to improve fertility ranging from identifying ovulation (even from spit, with startup companies they’ve featured like Ovul, a saliva-based fertility tracker), to the use of computer vision that helps researchers better identify which eggs and embryos will be most viable to have a woman carry to term. Some eggs would have been missed with the naked human eye, so this is truly a gamechanger!
tl;dr - Breaking down scientific jargon - AI model scans your image, takes encoded data, helps real live human embryologist grade and score your embryos are most viable for transfer, which improves success outcomes (what has the best likelihood to deliver babies).
This is hands-down this course is the most technically advanced I have seen on fertility sitting at the intersection of two key technologies, blockchain/web3/crypto (check out their AI bot, which was just released also this week), and all of the amazing founders they are funding to improve scientific research into fertility.
It was a privilege to hear Okta Chief Technology Officer, Bhawna Singh speak, sharing a wealth of insights about her career and all about AI and cybersecurity. We also learned a lot about cybersecurity and AI policy with a workshop presented by Maria Sokova, regulation as the space continues to evolve rapidly in the areas of data governance, privacy, trust, and safety. The bar for all the work we do in these complex areas is high. Luckily we have some of the sharp folks in the room working on these hard problems, creating solutions for the future. 🧠 I also got to be (re)acquainted with the venture investor, Eunice Wong, investor at Okta Ventures.
Big shout-out to The New Club's founder, Laura Du for organizing such a great space. For senior software engineers and technical women, consider applying to join the New Club. Let folks know I referred you!
AI - The DeepSeek vs OpenAI vs Everyone Else Debate
The gist: It is supposedly cheaper to use deepseek, so speculators will think that hardware companies like Nvidia would have less business—so you can see everyone from Meta (FAIR - Facebook AI Research) and their mom freaking out about this. At the same time, knowing Sam Altman, there are likely internal models ahead of the game compared to what DeepSeek has released, and others are skeptical. If you’ve been following X/Twitter and LinkedIn posts with every researcher, investor, seasoned AI practitioner or n00b, everyone seems to have a strong opinion about this whether or not they have tried deepseek, run their own evals (evaluation metrics), or they will be pro-America all the way, underestimating China.
What does this mean for the average consumer: Nothing (everything will be cheaper so that’s good for you), unless are super against using TikTok and all Chinese (or even Asian companies) and are pro-America all the way will you have a real issue with this. However, if you’re a software engineer, investor, or other business person you will care about two things:
1) Performance
2) Cost
If performance is as efficient with DeepSeek, but with cheaper cost you’ll go for #2. I remember a common saying by Jensen Huang (founder and CEO) of Nvidia,
“The more you buy, the more you save!”
As you read into why the founders of DeepSeek created what they did in the first place, you will find that they were unable to access even more H100s (Nvidia’s expensive Graphics Processing Unit - GPU), due to US export restrictions). DeepSeek ended up optimizing their stack more efficiently.
Should you believe everything you read? Is it that much faster and cheaper, are OpenAI and other U.S. based tech companies working on AI foundation/frontier models that much more behind? My advice referencing the web3 space: DYOR (Do Your Own Research). For machine learning practitioners like UC Berkeley PhD and startup founder, Josh Tobin (whom I had the privilege of taking Full Stack Deep Learning with a few years ago) said, “build your own evals.” Feel free to check out his great talk at MLOps Conference on “Evaluating LLM-Based Applications.” You can see a tutorial on how to do that here at OpenAI’s Cookbook.
Join our Taro paper reading group discussing the DeepSeek paper this upcoming week on the subject, see event details here
Taro is a YC-backed company which I love and highly recommend joining for those who are software engineers looking to level up their career.
Here’s a link to my QCon talk (some presentation slides I describe). InfoQ will release videos officially (right now it’s only available to QCon attendees unfortunately.
If you want me to speak at your upcoming conference, I am taking paid slots for this year. You can inquire and schedule me here.
Erin Jerri Malonzo Pañgilinan is a software engineer and computational designer. She is an internationally acclaimed author, publishing Book Authority’s #2 must-read book on Virtual Reality in 2019, O’Reilly Media book, Creating Augmented and Virtual Realities: Theory and Practice for Next-Generation of Spatial Computing, which has been translated into Chinese, Korean, and distributed in over 2 dozen countries.
She was also previously a fellow in the University of San Francisco (USF) Data Institute’s Deep Learning Program (2017-2018) and Data Ethics Inaugural Class (2020) through fast.ai.
She is currently working on her next books, applications, and films.
Erin earned her BA from the University of California, Berkeley. She is a proud Silicon Valley native.
