Hi Saloni! What is your advice/guidance for those who are interested in answering the call of medical research and innovation but don't know how? Over the past 3 years, I have frequently contemplated switching from software development to biology research, but really don't know how to make the transition. I have worked in the tech industry in a variety of software development roles for 17 years, so it feels easier to go towards building open source software tools and systems that biology researchers need/use, but it requires domain knowledge. So, I wonder if the best means to acquire this knowledge are traditional (university education), self-guided (online courses and specializations), or on the job (assuming there are roles where medical researchers prefer software systems engineers over computational biologists). Apologies if this isn't the right forum for my inquiry.
BTW, I love WIP physical magazines, and just received the August edition. Can't wait to read it. Thank you!
That's awesome – I'm working on a post with ideas about how to get involved, so will get back to you! Very glad to hear you're enjoying the magazine too.
Enjoying also your hard drugs podcast on various drug repurposing. You mentioned shingles vaccines and age given being quite old. I recently pushed my GP surgery for one whilst only 59 years old given I've had shingles twice within 5 years and it's getting worse. I got the good news and had my first last week with a follow-up in February next year. It's the inactivated one by the way. So people need to present at GP surgeries if they are getting shingles when younger multiple times and demand the vaccine you never know they might also get it on NHS like me.
The problem is that your solution operates in the same corrupt, money driven economy that we all are forced to live in.
The real solution is to put an AI in charge of research with an unlimited number of robot workers, who will work tirelessly for free. The AI and its workers can slave 24 x 7 and not pussyfoot about, working slowly, as they try to make their support money last as long as possible, which is what I believe happens with many research grants..
Like your example with the malaria vaccine, how many other discoveries have already been made but are languishing on the backshelves of laboratories? A search needs to be started.
And again, everything cannot be driven by money reward. There are 25-30 million cataract operations conducted annually across the world. These are operations where existing eye lens have to be replaced with artificial lens because even after decades and decades of study, no one in the whole wide world has yet been able to discover how to uncloud an eye lens or regrow an eye lens , so it doesn't have to be replaced at all. WHY NOT? Get an AI working on this!
The line that stuck with me is that the industry read the 1990s guidelines defensively, then kept doing the extra work even after regulators said it wasn't needed. A third of trial cost spent cross-checking data points the FDA itself says catch few errors is not a science problem. It's a process tuned to the wrong job.
Look at what each party is trying to get done. The sponsor wants to avoid rejection. The site wants to avoid audit findings. The patient wants to know, as fast as possible, whether this works for someone like them. Every step in today's trial process serves the first two, and the patient's job is treated as a side effect. RECOVERY and ring vaccination worked because someone started from the job of learning what works and built the process backward from there.
That's why I don't fully buy the low-hanging-fruit explanation for Eroom's law either. When tools get 10,000x better and output per dollar still falls, look at the system the tools are plugged into before blaming the frontier.
Hi Saloni! What is your advice/guidance for those who are interested in answering the call of medical research and innovation but don't know how? Over the past 3 years, I have frequently contemplated switching from software development to biology research, but really don't know how to make the transition. I have worked in the tech industry in a variety of software development roles for 17 years, so it feels easier to go towards building open source software tools and systems that biology researchers need/use, but it requires domain knowledge. So, I wonder if the best means to acquire this knowledge are traditional (university education), self-guided (online courses and specializations), or on the job (assuming there are roles where medical researchers prefer software systems engineers over computational biologists). Apologies if this isn't the right forum for my inquiry.
BTW, I love WIP physical magazines, and just received the August edition. Can't wait to read it. Thank you!
That's awesome – I'm working on a post with ideas about how to get involved, so will get back to you! Very glad to hear you're enjoying the magazine too.
Enjoying also your hard drugs podcast on various drug repurposing. You mentioned shingles vaccines and age given being quite old. I recently pushed my GP surgery for one whilst only 59 years old given I've had shingles twice within 5 years and it's getting worse. I got the good news and had my first last week with a follow-up in February next year. It's the inactivated one by the way. So people need to present at GP surgeries if they are getting shingles when younger multiple times and demand the vaccine you never know they might also get it on NHS like me.
The problem is that your solution operates in the same corrupt, money driven economy that we all are forced to live in.
The real solution is to put an AI in charge of research with an unlimited number of robot workers, who will work tirelessly for free. The AI and its workers can slave 24 x 7 and not pussyfoot about, working slowly, as they try to make their support money last as long as possible, which is what I believe happens with many research grants..
Like your example with the malaria vaccine, how many other discoveries have already been made but are languishing on the backshelves of laboratories? A search needs to be started.
And again, everything cannot be driven by money reward. There are 25-30 million cataract operations conducted annually across the world. These are operations where existing eye lens have to be replaced with artificial lens because even after decades and decades of study, no one in the whole wide world has yet been able to discover how to uncloud an eye lens or regrow an eye lens , so it doesn't have to be replaced at all. WHY NOT? Get an AI working on this!
The line that stuck with me is that the industry read the 1990s guidelines defensively, then kept doing the extra work even after regulators said it wasn't needed. A third of trial cost spent cross-checking data points the FDA itself says catch few errors is not a science problem. It's a process tuned to the wrong job.
Look at what each party is trying to get done. The sponsor wants to avoid rejection. The site wants to avoid audit findings. The patient wants to know, as fast as possible, whether this works for someone like them. Every step in today's trial process serves the first two, and the patient's job is treated as a side effect. RECOVERY and ring vaccination worked because someone started from the job of learning what works and built the process backward from there.
That's why I don't fully buy the low-hanging-fruit explanation for Eroom's law either. When tools get 10,000x better and output per dollar still falls, look at the system the tools are plugged into before blaming the frontier.
Empirical variance in these observed outcomes points to unmodeled environmental dependencies rather than intrinsic model failure.