Hype and Disruption: Post-Dagdigian Trends Talk Splits in Two
By Joe Stanganelli
May 7, 2024 | It finally happened. After being absent from last year’s “Trends from the Trenches” talk at the Bio-IT World Conference & Expo, BioTeam cofounder Chris Dagdigian has officially retired from delivering the traditional conference-closer.
Dagdigian’s charisma, rapid-fire delivery, and penchant for plain talk made the session a perennial favorite. Over the years, the session evolved from a solo presentation delivered by Dagdigian to a panel discussion. This year the session was merely bifurcated—split between a traditional-style “Trends from the Trenches” presentation by BioTeam CEO Ari Berman and a more focused talk delivered by BioTeam senior scientific consultant Laura Boykin Oklaebo. Below is a summary of the highlights:
Editor’s Note: Hear Berman’s Trends from the Trenches session on this month’s episode of the Trends from the Trenches podcast.
On Post-COVID Upheaval
Berman identified inconsistent and unaligned IT support as a major challenge facing scientists—a challenge exacerbated, Berman said, by the consequences of the COVID-19 pandemic. Complicated by a “re-differentiating of services” in high-performance computing leading to more options and more hardware specializations, Berman described hordes of quarantined researchers with no lab access suddenly independently deciding in tandem to analyze the piles of data they’d been sitting on. The result? VPNs were clogged with users clamoring for time, making on-prem systems hard to manage.
Meanwhile, pandemic-era economics have led to what Berman calls a “biotech winter.” Demand for lab space is down, investments from venture capital and private equity have dried up, and life-science organizations have become more sensitive to capital expenditures. “Innovation is slower,” said Berman. “No one was prepared for remote science.”
Berman reported that quarantines and shutdowns also led to a drastic spike in demand for cloud usage and cloud migrations that led to data-center operating costs spiking 10 to 50 times.
On Cloud
The bio-IT community has been increasingly dissatisfied with the state of cloud solutions for years. Highlighting both rising cloud demand and rising cloud hype, Berman called for “cloud sobriety”—criticizing the viewpoint of “If cloud, then all cloud.”
“One of the problems with humanity is that they tend to think in absolutes,” said Berman. “The reality is that [cloud] is a super-nuanced situation.”
On one hand, according to Berman, the cloud industry doesn’t deliver on modern marketing hype related to cost savings. He pointed to an issue that has come up in the Bio-IT World community before: that cloud fever has led to cloud customers effectively creating their own ransomware, being compelled to pay to put in their data and pay again to take it out.
“There’s a fundamental mismatch between the cloud business model and long-term research,” said Berman. “Saving things in the cloud can get really expensive [insofar as] we have to plan for infinite storage in our environments because no one throws anything away.”
On the other hand, going to the cloud offers customers undeniable benefits, including (but not limited to):
Large storage capacity
Automatic upgrades
Not having to worry about hardware lifecycles
Containerization and portability of workloads
Berman thereby urged that the solution is a hybrid-computing model preserving the user organization’s data ownership—while lamenting that cloud providers have yet to deliver on a sufficient hybrid model in this respect. “No one’s done it well yet,” said Berman.
On Working with Data
While the state of storage speed doesn’t particularly concern Berman (who advised that storage of 1PB+ doesn’t need to be especially fast, so long as it works), enterprise networks do. Berman argued that the oversubscribed, oversecured state of enterprise networks is slowing scientific research. Showing a slide with a picture of a FedEx truck, Berman quipped, “I’ve used this slide for so long; this is still the most common high-speed network.”
Berman also criticized the state of the FAIR data movement, noting that perverse individual incentives have stunted meaningful progress towards unity—i.e., that the more differentiated one is, the more likely they are to get funding. “How are we doing with FAIR? Pretty terribly,” said Berman. “This has to be a community effort, and there’s not a lot of incentive to be a community.”
On AI
In this, the last session of a conference that entirely devoted 30% of its tracks to AI plus several sessions outside those tracks, Berman gave a nod to the state of AI hype within the first few minutes of his presentation—showing a slide that said “AI” in large letters.
“AI. Thank you! We’re done!” joked Berman.
Returning to the subject more seriously later on, Berman suggested that AI interest among life-science organizations stems from dread over the seemingly insurmountable amount of data they’ve collected—and that the idea of falling back on AI to go through it all is comforting.
But the comfort may be premature. “Every organization—every person—tells me they need to be AI-ready; they don’t,” said Berman. “AI is not something we should 100% rely on… because it is still research.” Berman estimated that the life-science sector is still approximately two to five years away from AI productivity.
“In order to do some of these machine-learning algorithms, you have to have well-categorized, well-understood data,” said Berman. “So let’s pull back. Let’s be a little more realistic about AI. Let’s use it responsibly.”
On Access
After Berman’s presentation, Oklaebo related her experience helping cassava farmers in eastern Africa diagnose a virus strain killing their crops. With a portable genomics sequencer, Oklaebo and her team accomplished what might have been a six-month task in a matter of a few hours.
Oklaebo said the portable sequencer was necessary because of a lack of internet connectivity in the area her team was working in—par for the course in developing nations. Meanwhile, only one company—Oxford Nanopore—makes portable sequencers. Accordingly, Oklaebo focused her talk on calling for greater global access to technology.
“The diversity hotspots in the world are where there’s no cloud—and there are pathogens there,” said Oklaebo. “So if you don’t want to care because of equity, care because of human health.”
Oklaebo further argued that these considerations are all the weightier since the COVID-19 outbreak. “COVID exposed us,” said Oklaebo. “We’re going to think about it now, or we’re going to think about it later when the next thing comes out of a place [where] we don’t have resources.”