The Endpoints Slack interview: Saji Wickramasekara on where AI misses the mark — and what's next for Benchling

do you think the Y Combinator approach works in biotech today?

Most startups don’t work, whether in biopharma or software. Drug development is very hard and there is a huge benefit from experience. It almost looks like an apprenticeship business from the outside looking in. I do think that YC has encouraged more people to take a swing in bio and is helping reduce the barriers to entry. I also think there are non-biopharma applications of biotech that they have helped founders start companies in e.g. Solugen. From that perspective, I think their efforts are great.

are they still bio-curious today?

I’m 12 years removed from YC but from my point of view, they have always been motivated by working with great founders. Early articulations of a startup are usually terrible and evolve, so betting on the founders has always been their way.

Benchling, as an example, looked very silly 12 years ago and I hope we have proven that wrong 🙂

The biotech industry does need to do more to be approachable though. The pool of people who deeply understand science and software is very thin.

And the competition for tech talent is very different from anything you see in biopharma. But lots of talented people are interested in the mission vs. selling more ads.

Bezos and Altos?

Obviously a biotech success, but I don’t think some analyst put that in a spreadsheet for market size

Bezos and altos is cool though. Fully support rich people giving all their money to scientific research instead of buying more yachts.

To change gears to get a sense of what you’re seeing, how’d you summarize the biotech market vibe right now?

I feel like it’s been a pretty confusing year for most people to try to say where the market is

I’d sum it up as resolute and focused. Last year people were depressed. This year it feels like they are back to work. Science marches forward no matter what. There is confusion, but it’s coming from AI.

There’s a lot of work in front of us though. We still have grand aspirations of helping every biotechnology company get all their scientific data across research, development, and manufacturing, into one place. So all this AI talk actually comes true.

the interest in that technology must feel so different from when you started in 2012

That feels at the center of those smoke and mirrors — how much can these technologies actually be fully depended and relied upon with important work

Lots of attention is going to designing molecules but there’s so much toil in making medicines outside of that

Areas like these need more attention and could leap forward with AI

i think there’s this vague hope that the latest ChatGPT, with a bit of prompt engineering and tinkering, could get pretty close. but it feels like those tasks are so specialized and specific, that may be a dead end

My prediction…

So…short term we are overestimating, long term underestimating 🙂

Or like where we are with self-driving cars, it’s still a vast majority of humans driving around, even if there is some adoption.

Benchling was reportedly considering going public in late 2021. Why didn’t you?

Some of my sources speak, I think very aspirationally, of a window opening

If you have a few minutes still, let’s fire into a rapid round if that sounds good? Quickest thoughts here – I’ll take musings, fragments of sentences, emojis, whatever.

If you could speak to a young 22-year-old Saji when you were dropping out of MIT to start Benchling … what’s the one piece of advice you’d impart?

stay curious and keep a beginner’s mindset, being naive is good

Do you actually use ChatGPT or an LLM in your day-to-day life?

this is the real Saji, right?

ask me on monday

last Q: who should we invite next to the Slack room? This has been great!

I’ll let us return to our very real-life Fridays now. Thanks again for dropping by Saji!