Eliseo Papa
@eliseopapa.bsky.social
research engineering @deepmind science | previously #graph #ML for pharma | medicine @ImperialMed engineering @mit_hst
Reposted by Eliseo Papa
🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding.
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
March 25, 2025 at 5:25 PM
🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding.
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on the LM Arena leaderboard. 🥇
Proteasomes make natural antibiotics while breaking down proteins. So cool to see research in new areas revealing another layer of biological complexity. Makes you wonder what else is out there! www.nature.com/articles/d41...
Protein waste turned into antibiotics as a defence strategy of human cells
Proteasome cleavage changes during infection to generate antibiotics.
www.nature.com
March 10, 2025 at 10:28 AM
Proteasomes make natural antibiotics while breaking down proteins. So cool to see research in new areas revealing another layer of biological complexity. Makes you wonder what else is out there! www.nature.com/articles/d41...
Reposted by Eliseo Papa
nice perspective and a cool 🐟 study! www.nature.com/articles/d41...
Does a cell’s gene expression always reflect its function?
Transcriptionally similar neurons can be functionally diverse.
www.nature.com
February 14, 2025 at 7:41 PM
nice perspective and a cool 🐟 study! www.nature.com/articles/d41...
Reposted by Eliseo Papa
Delighted to have joined my good friend & colleague Noam Shazeer on a podcast with @dwarkesh.bsky.social for a 2+ hour discussion on early Google, ML hardware, training 1T+ token LLMs in '07, model sparsity, continual learning, and more.
Thanks, Noam and Dwarkesh! 🙏
youtu.be/v0gjI__RyCY?...
Thanks, Noam and Dwarkesh! 🙏
youtu.be/v0gjI__RyCY?...
Jeff Dean & Noam Shazeer – 25 years at Google: from PageRank to AGI
YouTube video by Dwarkesh Patel
youtu.be
February 12, 2025 at 9:59 PM
Delighted to have joined my good friend & colleague Noam Shazeer on a podcast with @dwarkesh.bsky.social for a 2+ hour discussion on early Google, ML hardware, training 1T+ token LLMs in '07, model sparsity, continual learning, and more.
Thanks, Noam and Dwarkesh! 🙏
youtu.be/v0gjI__RyCY?...
Thanks, Noam and Dwarkesh! 🙏
youtu.be/v0gjI__RyCY?...
Reposted by Eliseo Papa
Reading group session on Monday: Valentin De Bortoli joins us for his paper "Accelerated Diffusion Models via Speculative Sampling" arxiv.org/abs/2501.05370
On zoom at 9am PT / 12pm ET / 6pm CET: portal.valencelabs.com/logg
On zoom at 9am PT / 12pm ET / 6pm CET: portal.valencelabs.com/logg
February 1, 2025 at 7:38 PM
Reading group session on Monday: Valentin De Bortoli joins us for his paper "Accelerated Diffusion Models via Speculative Sampling" arxiv.org/abs/2501.05370
On zoom at 9am PT / 12pm ET / 6pm CET: portal.valencelabs.com/logg
On zoom at 9am PT / 12pm ET / 6pm CET: portal.valencelabs.com/logg
a call to all hardware founders to build josef.cn/blog/uk-talent in the uk 👏
UK's elite hardware talent is being wasted.
Imperial, Oxford, and Cambridge produce world-class engineers. Yet post-graduation, their trajectory is an economic tragedy - and a hidden arbitrage opportunity.
josef.cn
January 26, 2025 at 9:39 PM
a call to all hardware founders to build josef.cn/blog/uk-talent in the uk 👏
super cool paper!
“De novo designed proteins neutralize lethal snake venom toxin” paper from the Baker lab just published in Nature.
www.nature.com/articles/s41...
www.nature.com/articles/s41...
De novo designed proteins neutralize lethal snake venom toxins - Nature
Deep learning methods have been used to design proteins that can neutralize the effects of three-finger toxins found in snake venom, which could lead to the development of safer and more accessible an...
www.nature.com
January 17, 2025 at 1:40 PM
super cool paper!
We have been diving into the AMR world and it's fascinating! So many possibilities. Let's see what happens! 🧑🔬🧫🧪👨💻
www.fleminginitiative.org/post/ai-to-h...
www.fleminginitiative.org/post/ai-to-h...
The Fleming Initiative and Google DeepMind lay the path for AI to help tackle antimicrobial resistance
A new report from the Fleming Initiative and Google DeepMind lays out a vision for the impact that an AI and data-driven approach...
www.fleminginitiative.org
January 17, 2025 at 1:38 PM
We have been diving into the AMR world and it's fascinating! So many possibilities. Let's see what happens! 🧑🔬🧫🧪👨💻
www.fleminginitiative.org/post/ai-to-h...
www.fleminginitiative.org/post/ai-to-h...
Reposted by Eliseo Papa
A month ago we @vevotherapeutics.bsky.social announced that we have generated the largest single-cell perturbation atlas in history, Tahoe-100M. Today, we announce that we will fully open-source Tahoe-100M in Feb, as part of a collaboration with NVidia health to train cell state models.
January 13, 2025 at 4:23 PM
A month ago we @vevotherapeutics.bsky.social announced that we have generated the largest single-cell perturbation atlas in history, Tahoe-100M. Today, we announce that we will fully open-source Tahoe-100M in Feb, as part of a collaboration with NVidia health to train cell state models.
Reposted by Eliseo Papa
Interested in building the future of open source ligand- and structure-based ML models for ADMET prediction?
The ARPA-H funded OMSF (@omsf.bsky.social) OpenADMET project is hiring multiple positions to build ML models of ADMET properties and drive informative data collection!
openadmet.org/jobs/
The ARPA-H funded OMSF (@omsf.bsky.social) OpenADMET project is hiring multiple positions to build ML models of ADMET properties and drive informative data collection!
openadmet.org/jobs/
OpenADMET Jobs
Seeking talented scientists for the OpenADMET Consortium
openadmet.org
January 8, 2025 at 3:41 PM
Interested in building the future of open source ligand- and structure-based ML models for ADMET prediction?
The ARPA-H funded OMSF (@omsf.bsky.social) OpenADMET project is hiring multiple positions to build ML models of ADMET properties and drive informative data collection!
openadmet.org/jobs/
The ARPA-H funded OMSF (@omsf.bsky.social) OpenADMET project is hiring multiple positions to build ML models of ADMET properties and drive informative data collection!
openadmet.org/jobs/
Reposted by Eliseo Papa
Congrats to Johannes Linder, David Kelley et al. on the journal publication of Borzoi - a long context sequence models of RNA-seq coverage profiles with many nice applications for transcriptional & post-transcriptional regulation & variant effect prediction.
www.nature.com/articles/s41... 1/
www.nature.com/articles/s41... 1/
Predicting RNA-seq coverage from DNA sequence as a unifying model of gene regulation - Nature Genetics
Borzoi adapts the Enformer sequence-to-expression model to directly predict RNA-seq coverage, enabling the in-silico analysis of variant effects across multiple layers of gene regulation.
www.nature.com
January 8, 2025 at 10:06 PM
Congrats to Johannes Linder, David Kelley et al. on the journal publication of Borzoi - a long context sequence models of RNA-seq coverage profiles with many nice applications for transcriptional & post-transcriptional regulation & variant effect prediction.
www.nature.com/articles/s41... 1/
www.nature.com/articles/s41... 1/
Reposted by Eliseo Papa
This yearly list is fantastic as usual: medium.com/@tomwhitwell...
For example:
> To highlight tax evasion, South Korea introduced ugly neon green number plates for company cars worth more than $58,000. Luxury car sales fell 27%. [Song Jung-a]
For example:
> To highlight tax evasion, South Korea introduced ugly neon green number plates for company cars worth more than $58,000. Luxury car sales fell 27%. [Song Jung-a]
52 things I learned in 2024
This year I took a sabbatical, recorded a podcast with Freakonomics, designed and shipped a tiny modular synthesiser, and learned many learnings. Tom Whitwell designs open source music electronics as…
medium.com
December 19, 2024 at 3:12 PM
This yearly list is fantastic as usual: medium.com/@tomwhitwell...
For example:
> To highlight tax evasion, South Korea introduced ugly neon green number plates for company cars worth more than $58,000. Luxury car sales fell 27%. [Song Jung-a]
For example:
> To highlight tax evasion, South Korea introduced ugly neon green number plates for company cars worth more than $58,000. Luxury car sales fell 27%. [Song Jung-a]
Reposted by Eliseo Papa
#TeamTomo is at it again. Nice look at the conformational cycle of protein folding machinery inside native human cells by the teams of Frydman & Beck 🧪🔬🧶🧬
@nature, researchers report on their cryo-ET analysis of human TRiC/CCT -- the ATP-driven chaperonin the dysfunction of which is linked to disease. The work provides mechanistic insights into the functioning of this protein-folding machine within cells.
shorturl.at/H69Ye
shorturl.at/H69Ye
In situ analysis reveals the TRiC duty cycle and PDCD5 as an open-state cofactor - Nature
TRiC functions at near full occupancy to fold newly synthesized proteins inside cells.
shorturl.at
December 11, 2024 at 6:57 PM
#TeamTomo is at it again. Nice look at the conformational cycle of protein folding machinery inside native human cells by the teams of Frydman & Beck 🧪🔬🧶🧬
Reposted by Eliseo Papa
Turned @dippedrusk.bsky.social 's amazing Vancouver list dippedrusk.com/posts/2024-0... into Google Maps pins: maps.app.goo.gl/nGBbcUAMMixC...
Vagrant's Vancouver | Vagrant Gautam
A non-comprehensive list of places to go and things to do in the Greater Vancouver Area as curated by yours truly over 6 years. Might be outdated so please double-check!
dippedrusk.com
December 10, 2024 at 7:52 PM
Turned @dippedrusk.bsky.social 's amazing Vancouver list dippedrusk.com/posts/2024-0... into Google Maps pins: maps.app.goo.gl/nGBbcUAMMixC...
Reposted by Eliseo Papa
Lots coming for protein design!
December 8, 2024 at 10:33 AM
Lots coming for protein design!
Reposted by Eliseo Papa
If you will be at #NeurIPS2024 @neuripsconf.bsky.social and would like to come see our models in action, come say hi 👋 and check out our demo at the GDM booth!
Wednesday, Dec. 11th @ 9:30-10:00.
Lots of other great things to see as well! Check it out: 👇
deepmind.google/discover/blo...
Wednesday, Dec. 11th @ 9:30-10:00.
Lots of other great things to see as well! Check it out: 👇
deepmind.google/discover/blo...
December 6, 2024 at 12:42 PM
If you will be at #NeurIPS2024 @neuripsconf.bsky.social and would like to come see our models in action, come say hi 👋 and check out our demo at the GDM booth!
Wednesday, Dec. 11th @ 9:30-10:00.
Lots of other great things to see as well! Check it out: 👇
deepmind.google/discover/blo...
Wednesday, Dec. 11th @ 9:30-10:00.
Lots of other great things to see as well! Check it out: 👇
deepmind.google/discover/blo...
Reposted by Eliseo Papa
Excited to reveal Genie 2, our most capable foundation world model that, given a single prompt image, can generate an endless variety of action-controllable, playable 3D worlds. Fantastic cross-team effort by the Open-Endedness Team and many other teams at Google DeepMind! 🧞
Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents 🧠.
December 4, 2024 at 4:13 PM
Excited to reveal Genie 2, our most capable foundation world model that, given a single prompt image, can generate an endless variety of action-controllable, playable 3D worlds. Fantastic cross-team effort by the Open-Endedness Team and many other teams at Google DeepMind! 🧞
constructing good interviews is a never ending quest - here is a quick read on how access to LLMs changes incentives segfaulte.mataroa.blog/blog/hackerr...
Hackerrank was broken - but now it's actually harmful — segfaulte
segfaulte.mataroa.blog
December 4, 2024 at 10:08 AM
constructing good interviews is a never ending quest - here is a quick read on how access to LLMs changes incentives segfaulte.mataroa.blog/blog/hackerr...
Reposted by Eliseo Papa
as a kid, i fixed my parent's computer. as an adult I fix my kids's computer. pretty sure we are the only gen that knows how computers work
December 3, 2024 at 12:37 PM
as a kid, i fixed my parent's computer. as an adult I fix my kids's computer. pretty sure we are the only gen that knows how computers work
Reposted by Eliseo Papa
IMO VQGAN is why GANs deserve the NeurIPS test of time award. Suddenly our image representations were an order of magnitude more compact. Absolute game changer for generative modelling at scale, and the basis for latent diffusion models.
Taming Transformers for High-Resolution Image Synthesis
Designed to learn long-range interactions on sequential data, transformers continue to show state-of-the-art results on a wide variety of tasks. In contrast to CNNs, they contain no inductive bias tha...
arxiv.org
November 28, 2024 at 12:09 AM
IMO VQGAN is why GANs deserve the NeurIPS test of time award. Suddenly our image representations were an order of magnitude more compact. Absolute game changer for generative modelling at scale, and the basis for latent diffusion models.
Reposted by Eliseo Papa
MMseqs2 Release 16 Highlights: GPU-accelerated search📄, ORF or new 6-frame translated search modes, contig taxonomy always keeps the longest ORF, bug fixes (reduced memory and higher sensitivity) and relicensed as MIT
📄 biorxiv.org/content/10.1...
💾 mmseqs.com and 🐍Bioconda 🖥️🧬🧶
📄 biorxiv.org/content/10.1...
💾 mmseqs.com and 🐍Bioconda 🖥️🧬🧶
November 27, 2024 at 9:08 AM
MMseqs2 Release 16 Highlights: GPU-accelerated search📄, ORF or new 6-frame translated search modes, contig taxonomy always keeps the longest ORF, bug fixes (reduced memory and higher sensitivity) and relicensed as MIT
📄 biorxiv.org/content/10.1...
💾 mmseqs.com and 🐍Bioconda 🖥️🧬🧶
📄 biorxiv.org/content/10.1...
💾 mmseqs.com and 🐍Bioconda 🖥️🧬🧶
Reposted by Eliseo Papa
What BlueSky got right that Mastodon got wrong. For the 99% of users that don’t give a FF about federation, it works exactly like old Twitter, and they never need to know. Classic Apple-epitomized design principle: Advanced features should be easy to access but, when unwanted, disappear entirely.
November 27, 2024 at 2:23 AM
What BlueSky got right that Mastodon got wrong. For the 99% of users that don’t give a FF about federation, it works exactly like old Twitter, and they never need to know. Classic Apple-epitomized design principle: Advanced features should be easy to access but, when unwanted, disappear entirely.
microbiome perturbation in the gut mapped spatially, how cool
1/ Keen to share the story of my postdoc research out in @natureportfolio.bsky.social! We built a massive resource characterizing the spatial landscape of the gut—and yes, we dug deep into the biology too. If you’ve got a minute, let me walk you through the story. 🧵👇
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Spatially restricted immune and microbiota-driven adaptation of the gut - Nature
Intestinal regionalization is characterized by robust and resilient structural cell states and the intestine can adapt to environmental stress in a spatially controlled manner through crosstalk betwee...
www.nature.com
November 22, 2024 at 10:05 AM
microbiome perturbation in the gut mapped spatially, how cool