venu☘️

🍳 techbio

[venu☘️ @itsvenu_]

because teams don't talk to each other

[rahul @0interestrates]

why does google have two competing coding agents - antigravity and jules?

[venu☘️ @itsvenu_]

it's wild out there

[venu☘️ @itsvenu_]

people posting about the interview process for andrej to join anthropic.

i hope they are talking about andrej interviewing anthropic, not the other way around, right?

[Andrej Karpathy @karpathy]

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

[Pubity @pubity]

Taiwan scientists have created a new serum that regrows hair in just 20 days

[Nathan Benaich @nathanbenaich]

It was such a pleasure to spend a Saturday morning with @RecursionChris recording this pod in NYC to talk about the @stateofai and biotech in 2025🦎 👀

[venu☘️ @itsvenu_]

asking the right questions at molecular level is really critical in understanding ageing.

even more important is, answering these questions with currently available means, data, technology. @Sanjusinha7 lab has done a really good job at this.

[Sanju Sinha @Sanjusinha7]

When and how do different tissue physical structures deteriorate during aging (structural aging)? What molecular changes occur in tissues during periods of major changes?

Are there tissue-specific periods of accelerated structural aging? Which organs age early vs. late?

Is there cross-organ coordination in structural aging? And, finally, how lifestyle, diseases and genetics impact these organ-specific trajectories?

While important, these questions haven’t been answered yet because we lack structural and molecular data from normal aging tissues at scale.

We present a framework taking the first stab at scale at these questions using high resolution histology images and omics (+more) from 25,000 post-mortem tissues (public data: GTex).

We reason that structure determines function and learning how tissue structure changes with age can help us understand the process of aging in different tissues - a central question with yet little understanding. An example is to visualize these two ovaries histology: young vs. old ovary- young ovaries cortex is intact, with follicles, no fibrosis - basis of its function (partially).

First, we extract tissue structure from these organs using their high-res histology images using a pre-trained digpath foundation model (UNI) and asked how much they change with age (Structural Aging Rate)?
As an example, the ovary has two peaks around the late 30s and then around 55. First aligns with accelerated follicle loss and second is just after post-menopause.

This shows that change in morphology of the ovary captures its functional milestones during aging with no training.

Bonus Puzzle: Does anyone know how these two functional milestones of ovary were originally found in the last century?

What if we repeat the same analysis on bulk-omics, transcriptomics and methylation, from the same samples? Can they capture this bimodal functional decline? No. (read the paper of our explanation)

Can molecular clocks trained on chronological age track this? No. They assume aging is linear.

Note: Unlike molecular clocks trained on chronological age, PathStAR learns with no age labels. We simply ask: When and how does tissue morphology change most rapidly during life?

[NotebookLM @NotebookLM]

We're partnering with @GoogleResearch to kick off a new series of Featured Notebooks exploring topics ranging from health to science to (of course) AI.

Starting with "How do scientists link genetics to health" this notebook explores some of what we've learned over the past 10 years of genomic research:

https://t.co/XhI2XgNlhj

[Thinking Machines @thinkymachines]

Introducing Tinker: a flexible API for fine-tuning language models.

Write training loops in Python on your laptop; we'll run them on distributed GPUs.

Private beta starts today. We can't wait to see what researchers and developers build with cutting-edge open models!

https://t.co/tJsgxgBuWo

[Anand Bhattad @anand_bhattad]

So You Want to Be an Academic? A couple of years into your PhD, but wondering: "Am I doing this right?" Most of the advice is aimed at graduating students. But there's far less for junior folks who are still finding their academic path.

My candid takes: https://t.co/25JdxHAON0

[venu☘️ @itsvenu_]

because they didn't survive to get it third time?

[Chief Nerd @TheChiefNerd]

ROGAN: “Did you get COVID?”

JENSEN: “I got COVID like six times.”

ROGAN: “Of course you did, because you got vaccinated. The people who didn’t get vaccinated got it once or twice.”

JENSEN: “Really? … Sh*t!”

[venu☘️ @itsvenu_]

chat is knowing me better and better each day

[venu☘️ @itsvenu_]

chat ❤️

[venu☘️ @itsvenu_]

https://t.co/AMv3xiRUpP

[venu☘️ @itsvenu_]

@OpenAI

[venu☘️ @itsvenu_]

people don't know what they need, you need to tell them what they need and sell that to them — steve jobs(!)?

[Garry Tan @garrytan]

I don’t understand who asked for a thinner iPhone

[venu☘️ @itsvenu_]

@nathanbenaich schnitzel in vienna is as big as vienna itself :)

[venu☘️ @itsvenu_]

sure, whatever you say

[Claude @claudeai]

We’d also like to address a concern we've heard in the community: We never intentionally degrade model quality as a result of demand or other factors.

[Daniel Gerlach @gerlach_d]

Our new work is now published in @CellRepMed: Combining BET inhibition with SMAC mimetics restricts tumor growth and triggers immune surveillance in preclinical cancer models (https://t.co/a6Dl6XBDws)

BI 894999 available: https://t.co/fZKCDhBrEg

#OpenAccess #BETi #SMACmimetics

[venu☘️ @itsvenu_]

killing all language learning apps with an 'update'. GOAT 🐐

[Google @Google]

🧵 We’re announcing two new updates to Google Translate to make it easier to connect with people who speak different languages, using the advanced reasoning and multimodal capabilities of Gemini models.

First: Starting today in the Translate app, you can tap “Live translate” and have a back-and-forth convo in real time with audio and on-screen translations — in more than 70 languages. 🗣️

Available this week for users in the U.S., India and Mexico.

[venu☘️ @itsvenu_]

@maharshii