Peter Ebasone (@VanesPee) header
Peter Ebasone (@VanesPee) profile picture

MD, PhD — Computational Bio & Precision Medicine.

Joined September 2010
Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)Photo by Peter Ebasone (@VanesPee)
AI and machine learning: learn it now or regret in a few decades.
1
2
7
About a year ago, we launched @useaccrue in Cameroon, and today we’re almost at 14,000 customers there. I’ve never been to Cameroon. When we started, I didn’t know anybody there or anyone who knew anybody there, and I didn’t speak French. It started on our dashboards. Cameroonians were signing up for Accrue, and we had done no marketing there and had no employee there. Around the same time, one of my co-founders posted in our group chat that a lot of Cameroonians import goods from Nigeria to sell in their markets. We suspected there was something for us there, but I needed to talk to people to be sure. So I paid for LinkedIn Premium and cold messaged marketers, salespeople and 9-to-5 workers in Cameroon. Some of the messages I translated into French with ChatGPT. A bunch of people replied, and I got on calls with them. The common thread was that they would need something like this, and there was no option like it in Cameroon. That gave my suspicion legs to stand on. I couldn’t do it from Nigeria, so we hired someone on the ground through LinkedIn. She had some marketing and sales experience, but for the most part it was a gut feeling. She’s been amazing. In Cameroon, sending money to Nigeria usually means using a guy who knows a guy with a Nigerian bank account, at an exchange rate you don’t like. There aren’t many fintechs there, so a lot of education had to come into play, and some of it can’t happen online. Our lead in Cameroon hired agents who go from place to place, explaining Accrue to people and signing them up. That’s how we got to almost 14,000 customers, and it keeps growing month over month. I also didn’t bargain for the language barrier. Calls that should take 30 minutes were taking an hour and a half because someone had to interpret. So in March, I got a Cameroonian French teacher. I now speak bad French very well, and she gets paid every month on Accrue. I used to think people doing interesting things just know for sure it will work. Every time something works for us, I’m reminded that you actually don’t know. You’re just messing around and finding out. Before the gain, there will be a lot of losses. Be fine with that. I wrote the full story on Medium: medium.com/@adesuwaomoruyi/h…
80
178
23
1,137
33,751
Peter Ebasone retweeted
A local guide show how to deal with quicksand during a tour of Mont Saint Michel.
585
4,415
584
57,556
27,421,262
Peter Ebasone retweeted
"My kids listen to me when I ask them to do stuff. Parenting has just become so much easier. When we are around other parents and they have rules around candy, screen time, etc. it all just seems so exhausting enforcing those rules. I'm really happy with the outcome we have gotten from this book." Gratifying: reddit.com/r/homeschool/s/L6… "I read the book a year ago and implemented many of the concepts in our house and it has been a game changer. I'm honestly surprised that more people aren't doing it. It seems radical but out of all the parenting books I've read this one has had the most positive impact. I have a 7, 6, and 2 year old. Our kids have zero restrictions on food, screen time, or bedtimes and they are all thriving. My oldest is a very strong willed and highly sensitive child. She absolutely hates being told what to do and we were having massive fights on trying to get her to do as we say. All most all of that has gone away and she is a delightful and easy child now. My kids eat and drink whatever they want. I don't keep sodas in our house but when we go out they can have whatever soda they want and their behavior remains the same. They eat a variety of fruits and vegetables along with cereals and frozen pancakes. The other day my two year old wanted ice cream for breakfast so I gave it to him. Later he ate eggs and toast. Food just isn't an issue in our house and it's so much easier! My kids willing share candy with each other without being asked. If they ask for candy I give them whatever we have whenever they want. My only rule is that I ask them to try food that I cook for them. I explain to them that it takes a lot of work to make them food and it's polite to try it. For the most part they will try whatever I cook, but I always try and make them foods that they like. If I feel like having a grown up meal then I'll make them chicken nuggets along with our meal. The unlimited screen time was an adjustment. At the beginning my kids clocked 10 hours on average a day on their iPads, mostly playing Roblox and watching YouTube. My daughter basically taught herself to read because she wanted to read the Roblox chat. She's now in advanced reading. After a year of unlimited screen time they now are bored with their iPads and they want to go outside and play, play card games with each other, the screens don't hold their attentions like they use to. And guess what, their attention spans and behavior are just fine. And they never ask us for more screen time. Screen time isn't an issue and I feel like their relationship with it is healthy. For bed time we all go to bed around the same time. If they want to stay up late they are more than welcome to but they mostly go to bed with us. My two year old will even come tell me when he's ready for bed. Bedtime isn't a struggle and my kids aren't overly tired. My husband and I don't get much alone time together but that's just the phase of life we're in. I'm honestly surprised that more people haven't taken to this approach. My kids are so much happier. My two year old rarely has tantrums. My kids are so kind and share everything with him. People often compliment me on how mature, kind and patient my kids are with other little kids. My kids listen to me when I ask them to do stuff. Parenting has just become so much easier. When we are around other parents and they have rules around candy, screen time, etc. it all just seems so exhausting enforcing those rules. I'm really happy with the outcome we have gotten from this book."
5
6
1
57
4,334
Peter Ebasone retweeted
After an explanation lands, learning can be extremely fast. Before an explanation lands, all the typical "virtues" are required: hard work, willpower, resilience, and self-discipline to endure drudgery.
Four kids now from age 12 to 2. My biggest regret is underestimating how rapidly they can develop skills with even 5-15 minutes a day of focused effort. School makes you think a reasonable learning period is one hour. That is way too long. Even 5 minutes stacks up quickly.
3
6
48
2,151
Peter Ebasone retweeted
A full generation of young scientists have grown up thinking the p-value is the way to measure scientific rigor. That is a collective failure of our scientific training mission. It's much easier to teach, but it has had a very negative effect on how young scientists think.
37
76
15
583
54,163
Peter Ebasone retweeted
How multimodal AI, combined with blood test omics via "liquid biopsies," will transform cancer monitoring and enable earliest detection. open-access @CellCellPress cell.com/cell/fulltext/S0092…
12
62
2
233
25,285
Peter Ebasone retweeted
Together with e-means, we are officially releasing the first 10 billion scale benchmark for clustering and search : Dino10B. It consists of 10 billion embeddings from Yahoo Flickr image patches computed with a Dino ViT-L/16 model. Blog post at : swannbessa.github.io/blog/di…
11
37
1
387
12,825
classic ML ftw! 🔥
Introducing e-means, an algorithm beating k-means in 10x fewer epochs. With order of magnitude wall-clock speedups, e-means is released in FAISS and needs no hyperparameter tuning beyond k. It is robust to various scales and datasets. Still skeptical ? Just try e-means [1/10]🧵
1
85
Peter Ebasone retweeted
Two of the most important opportunities for medical AI: 1. Give the gift of time to clinicians, to promote the patient-doctor relationship 2. Improve accuracy of diagnosis and treatments Both are helped by this type of work
Excited for the launch of Gemini for Healthcare! We’re extending the Gemini agent with specialized healthcare skills and connectors to Epic, Optum, and more, built to support HIPAA compliance. Instead of spending hours piecing together fragmented records and policies, care teams can offload the admin burden and get time back for what matters most: patient care.
32
38
7
283
28,321
Single-cell eQTL mapping just got better. SAIGE-QTL works directly with single-cell counts instead of averaging cells first, and found 48.8% more genes with eQTLs in 1.2M immune cells. nature.com/articles/s41588-0…
1
64
Peter Ebasone retweeted
We won the Democratic trifecta — now it's time to get to work.
83
111
3
775
16,526
a lot of people speak with the kind of confidence you can only get by never having done or having to do the work
11
358
32
846
68,296
Interesting preprint: PALM fine-tunes ESM2 to predict protein domains and functional sites. Without ever training on ClinVar, its zero-shot score (paired with ESM2) beats AlphaMissense on 128k literature variants outside ClinVar. biorxiv.org/content/10.64898…
1
86
Peter Ebasone retweeted
People who reason by first tracing things end-to-end have a very different process that can look particularly stupid until it very, very suddenly clicks in to a creepy level.
1
2
22
911
Your genetic risk variant may have been harmless for 60 years. Then the aging RNA machinery stops buffering it. New review in Trends in Cell Biology by our grad student @sgrandles : RNA dysregulation belongs alongside DNA damage & proteostasis as a driver of neuronal vulnerability. sciencedirect.com/science/ar…
6
11
32
2,342
if you’re unsure about where you want to go in life, an easy step forwards is to figure out where you absolutely don’t want to be and get away from there.
2
120
New in @NatStructMolBio! 🧬 We screened 730 RNA-binding proteins and found 281 that control circular RNAs, from how they’re made to how much protein they produce. Thanks to our collaborators at MSK, Columbia, and UCSD, and to my amazing lab, led by Tao Yu nature.com/articles/s41594-0…
2
33
124
7,381
Peter Ebasone retweeted
BE Seminar this Thursday with @arjunrajlab @PennEngineers “Systems of Cells and Systems of Science” ow.ly/b7Yl5103agn
5
25
1,932
Interesting paper: Fifteen challenges for generative AI applications to cell biology sciencedirect.com/science/ar…
39