deepakalur (@deepakalur) header
deepakalur (@deepakalur) profile picture

builder | investor | advisor | Ex @Amazon @OneMedical @OpenGov @Anaplan @Sun. into Health tech. opinions my own, changes mind with learning. RT≠ endorsement

Oakland, CA
Joined October 2008
Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)Photo by deepakalur (@deepakalur)
One interpretation: jobs that are complex human-facing will increase, jobs that are procedural and algorithmic will decrease. Most patients, i.e. 100%, still want a human in the loop in these domains.
New essay - healthcare jobs are basically keeping the US labor market afloat, but it's starting to see some weakness. This raises questions about how local economies are going to prepare. There are three big factors I see: 1) The government is now cracking down on how money is getting pumped into states. I took a look at the data and governmental transfers to states via Medicare/Medicaid are now an enormous amount of dollars in that economy. What happens when that gets shut off? 2) Local politicians are starting to realize that the healthcare jobs program is actually causing non-healthcare jobs to avoid the state. Indiana provides an interesting example here. 3) AI is going to hollow out many admin jobs. I sliced the BLS data which actually flags which jobs are at risk for automation. It also presents an issue where full automation will extract dollars from these economies and concentrate them where the AI companies are based (e.g. SF, NY, etc.).
25
From $2T down to $1.3T valuation is heading in the right direction - a 40% drop in a couple weeks (if you had that stock in the secondary market!). The next phase of this IPO will be more interesting...
AI euphoria could be coming to an end
30
"I believe that the company which gets the risk right will also be the company that has the license to operate...We have to make sure that all the risks, all the ways that things can go wrong with this technology, we very carefully address those risks." But, hey let's not wait to IPO! A touch of arrogance, a pinch of delusion, and a lot of chest thumping mixed in. ACCELERATE SAFETY
Anthropic CEO predicts cancer, Alzheimer's, heart disease, mental illness will be cured in less than 10 years ACCELERATE
1
15
interesting. thoughts?
AI companies admit what they’re building might kill everyone. We don't have to agree to this. Add your name at donotaccept.ai to say: we do not accept your terms.
1
14
way to go!
You shouldn’t have to sell 7% of your company to join a top accelerator. That's why FC Build is completely free. We take zero equity and charge zero fees. And we’re currently recruiting for our next cohort of technical founders at the earliest stages. You get: → $1M+ in startup credits → Direct access to our investment team → A small, highly selective cohort of exceptional founders → Hands-on support, workshops, office hours, and SF office space Interviews for FC Build are already underway, don’t wait to apply. Applications close November 20: foundation.typeform.com/fcbu…
20
“Since Silicon Valley pivoted to Trump, then, its scorecard has looked something like this: chief rival (China) strengthened; domestic support weakened; global markets infuriated.”
Getting too close to Trump was a mistake – how could Silicon Valley have gotten (US) / got (UK) it so wrong? Read the full article: theatlantic.com/ideas/2026/1…
18
so i just got access to @claudeai duet and signed in - i haven't done anything in it yet, but it already shows 12% used. what gives?
1
80
I am really grateful to @hnshah for all the good things he puts out there to share his knowledge and expertise - for free and for fun (he says)! Follow him to find out. I just attended his session "Coding Agents 101: For People Who Don’t Write Code" - there was a time...oh maybe 6 months ago where i thought i would always write code again, but the last 2 months have convinced me that i never will write code myself again. I also stumbled on his treasure trove of a library here for you to watch/learn hiten.com/library - Thanks Hiten!
1
1
102
it was always files
The future is files. And GPUs.
22
deepakalur retweeted
We have 80 empty desks in our SF SoMa office, and we’d rather fill them with founders. Free for early-stage teams of 2–20. 6 months. No equity, no strings. Come build next to us at Pump.co 👇 forms.gle/92wKAja5mkMdXSQ47
68
37
10
513
56,720
😆
How can you say AI kills creativity ? This is amazing.
18
deepakalur retweeted
Actually very useful 👇
If you get a massive surprise medical bill from a US hospital: ABSOLUTELY AVOID THESE - Paying it immediately out of panic. - Setting up a payment plan (this legally admits you owe the full amount). - Calling the billing department to argue on the phone. - Believing "insurance denied it" means you have to pay. IF YOU WANT TO SLASH THAT BILL FROM NOW ON ↓
3
87
38,030
ballot props are crazy. anyone can float one with enough signatures, write it deceptively so that voters who read the 2 lines on the ballot feel like supporting it. not many read the details. The state law makers have to look into this proposition.
This is Gary Michelson. Spent around $72 million to get prop 38 on the ballot. Now a non-profit he controls will get $4.2 billion from taxpayers. The prop also includes statutory exemption for his non profit from public sector transparency and ethics laws. Generational grift. Ballot props are a bad idea.
1
1
42
Great talk. Everyone working on this space should watch this. Glad to Jay Shaw summarize and emphasize on data foundation, which IMO is the most important thing for AI to work well. Also - bring the patient along, explain how you are using their data for their own better outcomes. Don't be creepy, don't try be magical without explaining.
How can we responsibly use AI to predict a patient’s future? Jay Shah, MD from @concurrencehq, and @Dr_PJay, MD, PhD from Optum and Dell Medical School joined us at the Second Opinion Summit to discuss how decision intelligence, strong data foundations, and patient trust will shape the next era of healthcare. Tune in to the full Lifers episode: youtube.com/watch?v=7FSxZYR7…
20
It has never been not ok to do that. It is indeed Artificial Intelligence, always will be from a human perspective.
Is calling it "artificial intelligence" okay again?
25
deepakalur retweeted
My mom passed away from cancer when I was eight. Both of my parents were doctors, and I've often thought about what my family went through during those years. Peter is helping his mom fight cancer. Along the way, he's built a free AI skill that helps families navigate the overwhelming amount of medical information and prepare for the decisions ahead. I find it incredible that this kind of help is now possible. Even more incredible that Peter is making what he's built available to anyone who needs it. This is one of the most meaningful uses of AI I've seen. github.com/petergyang/fuck-c…
Since I shared that I was helping my mom fight cancer, I've learned a lot about what a caregiver can do. Here's my new video where I walk through 5 steps that I'm taking and what I wish I'd known sooner. The sad reality is that 2 in 5 people in the US will get cancer in their lifetime. If you include caregivers, pretty much every family is impacted. This is my attempt to share what I've learned with anyone going through the same thing. I hope you find it useful. 📌 Watch now: youtu.be/Ddc5NFNnCfE
22
53
2
551
27,315
as predicted Q3 numbers aren’t keeping up. interesting what the valuation will be considering this and other factors that have changed since they put out the S-1
want to see Q3 numbers. Q1/Q2 were heavy because everyone was tokenmaxxing. Also the cut prices in half, what’s the impact of that on revenue and market share. roadshow will be interesting
20
Looking at all election results I can’t help but feel that Citizens are united against Citizens United; all that trillionaire / billionaire money didn’t seem to dent the democratic process this time. The American Experiment is still on. #BlueTsunami2026
22
Sandboxes have no moat. Just fat margins. We're launching agent37.com to kill those margins. For example, a sandbox with 2 vCPUs and 4 GB RAM when left running for an entire month: Vercel: $249 Cloudflare: $145 Daytona: $121 Agent37: $4.76 There's nothing magical about a sandbox that justifies keeping it expensive forever. Call it a race to the bottom. We'll see you there. If you can find a cheaper sandbox and we'll refund the difference.
Readers added context they thought people might want to know
Vercel and Cloudflare charge for active CPU time (excluding idle/waiting periods common in agent use) and provisioned memory while running, unlike the full-month 100% utilization figures quoted; Daytona bills full reserved resources while started. vercel.com/docs/sandbox/p… developers.cloudflare.com/containers/pla… daytona.io/pricing agent37.com
404
111
181
1,749
1,405,154
spy vs spy between these two people makes for good entertainment
Replying to @unclebobmartin
“Your job is to help the LLM write better code faster.” Strange: I thought the job of a software engineer was to bring value to stakeholders in an enduring, economically sustainable, and ethical fashion. I suggest you reconsider your life choices, Bob.
29