#AI interpretability

[mlobsters]

jared padalecki and jensen ackles boston con 2025 - gold panel

cheerleading and (allegedly non-existent) short shorts

[jabba-the-fett]

“We make our own decisions. Our own choices. And we have to live with them too” - Commander Cody

Idk how digital art works, but here’s a Commander Cody I spent way too much time on

[wackywatchdotcom]

i think theoretically the idea of caine adding in some sort of pet would be very fun and cute but i cant help but think pomni would be unnerved by it

[pharawee]

"How can it be your fault? Let's just let fate runs its course, whether it's your fate, my fate our both of our fates combined."

—I SAW YOU IN MY DREAM · Episode 11

[falsehydra101]

Thetis if Mythology was based (She is also my OC now I suppose✨)

[osteochondraldefect]

Sweet reward for obeying commands

[snoopybleedstoo]

Bloody Judge

Judge Holden portrait from the book "Blood Meridian" I started this digital painting sober and then later didn't

[aloneeyez]

Sillys

[holydivers]

like look i'm as anti-ai as the next guy but it seems super disingenuous to say "it's absolutely inexcusable to cause so much environmental damage and waste so much water just for a tiny bit of convenience or amusement that's unethical on top of that" okay. so are you giving up beef (or, heaven forbid, meat and dairy as a whole) or do you only actually give a shit about making sacrifices for the sake of environment when it just happens to be a thing you already don't like?

just be honest and say you hate it bc it sucks and it's ugly. because it is

edit: also vegans who use AI i am going to attack you with hammers. "oh but this ai comes up with veganized versions of recipes!" ok. either a) it's just getting that info from a recipe that someone already made, in which case just go find that fucking recipe yourself. or b) it's just making shit up, which you can just do yourself by looking at substitutes for existing ingredients. the ai doesn't fucking know if aquafaba will work in a certain recipe. get off the stupid little app and trial and error it in the kitchen the way it's meant to be

[jcmarchi]

TheSequence Radar #674: Transformers in the Genome: How AlphaGenome Reimagines AI-Driven Genomics

New Post has been published on https://thedigitalinsider.com/thesequence-radar-674-transformers-in-the-genome-how-alphagenome-reimagines-ai-driven-genomics/

TheSequence Radar #674: Transformers in the Genome: How AlphaGenome Reimagines AI-Driven Genomics

A model that could advance the future genomics.

Created Using GPT-4o

Next Week in The Sequence:

Knowledge: An intro to the world of multi-agent benchmarks.

Engineering: Let’s hack with the Gemini CLI.

Opinion: Why circuits could be the answer to AI interpretability?

Research: AlphaGenome deep dive.

Let’s Go! You can subscribe to The Sequence below:

TheSequence is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

📝 Editorial: Transformers in the Genome: How AlphaGenome Reimagines AI-Driven Genomics

I have been obsessed with AI in genetics for some time so I couldn’t write about anything else today other than DeepMind’s new model: AlphaGenome!

AlphaGenome merges some of the best -established techiques in AI-driven genomics such as large-scale sequence context with base-pair precision to chart the regulatory genome in a way never before possible. The model’s four-headed architecture digests up to one million contiguous base pairs in a single pass, outputting synchronized predictions for chromatin accessibility, transcription-factor occupancy, RNA expression, splicing, and 3D genome architecture. This unified approach replaces fragmented, single-modality pipelines—each requiring separate models and datasets—with one cohesive model that excels across tasks, streamlining variant effect analysis for researchers.

At its core, AlphaGenome marries convolutional layers, which capture local nucleotide motifs analogous to transcription-factor binding sites, with transformer modules that integrate distal regulatory elements hundreds of kilobases apart. DeepMind’s design eschews downsampling, ensuring every nucleotide contributes to high-resolution inferences. As functional genomics datasets from consortia like ENCODE, GTEx, and 4D Nucleome expand, this backbone stands ready to unveil regulatory grammar buried deep in non-coding DNA.

Traditional genomics models often excel at one signal—SpliceAI for splicing, ChromBPNet for chromatin state—necessitating an ensemble of tools to profile variant consequences fully. AlphaGenome’s simultaneous four-headed predictions eliminate this bottleneck, revealing cross-modal interactions—e.g., how a variant that disrupts a splice site may also alter local chromatin loops—opening novel avenues for mechanistic insight.

In benchmark evaluations spanning 24 sequence-prediction and 26 variant-effect tasks, AlphaGenome matches or surpasses specialized baselines in over 90% of cases. It outperforms SpliceAI, ChromBPNet, and other state-of-the-art models by significant margins, all while completing variant-effect scans in under a second—transforming in silico hypothesis testing from minutes or hours to real-time speed.

The genomics market in 2025 stands at an inflection point: cloud-based sequencing costs have halved over five years, single-cell and long-read technologies have become routine, and multi-omic datasets proliferate. Yet, analytical bottlenecks limit the translation of raw data into actionable insights. AlphaGenome arrives precisely when biotechnology and pharmaceutical companies demand scalable, AI-driven interpretation to bridge the gap from variant discovery to biological understanding. Its ability to standardize and accelerate regulatory variant annotation is poised to catalyze next-generation diagnostic tools, precision therapeutics, and synthetic biology platforms, redefining competitive advantage in a data-saturated market.

DeepMind’s preview API grants non-commercial researchers early access to AlphaGenome, democratizing large-scale regulatory predictions. From pinpointing causal non-coding mutations in disease cohorts to engineering synthetic enhancers with bespoke cell-type specificity, this open sandbox invites collaborative breakthroughs across academia and industry.

If AlphaFold decoded protein structures, AlphaGenome now deciphers the regulatory code—the “dark matter” governing gene expression. As single-cell, long-read, and cross-species datasets proliferate, the model’s extensible architecture promises seamless integration of new modalities. The future of genomics is computational, and AlphaGenome lights the path forward: an intellectual and technological leap toward understanding—and ultimately rewriting—the language of life.

🔎 AI Research

AlphaGenome: advancing regulatory variant effect prediction with a unified DNA sequence model

AI Lab: Google DeepMind Summary: AlphaGenome is a deep learning–based sequence-to-function model that ingests one megabase of DNA sequence and predicts thousands of functional genomic tracks—including gene expression, transcription initiation, chromatin accessibility, histone modifications, transcription factor binding, chromatin contact maps, and splicing—at single-base-pair resolution. Trained on both human and mouse experimental data, it unifies long-range sequence context with high prediction resolution, outperforming prior methods and enabling comprehensive in silico characterization of regulatory variant effects.

Confidential Inference Systems: Design Principles and Security Risks

AI Lab: Pattern Labs / Anthropic Summary: This whitepaper defines the architecture of a “confidential inference system” that leverages hardware-based Trusted Execution Environments (TEEs) to protect both user data (model inputs/outputs) and model assets (weights and architecture) during AI inference workloads. It further details reference designs for secure model provisioning, enclave build environments, service provider guarantees, and a comprehensive threat model to mitigate systemic and implementation-introduced risks.

USAD: Universal Speech and Audio Representation via Distillation

AI Lab: MIT CSAIL Summary: USAD distills knowledge from multiple domain-specific self-supervised audio models into a single student network capable of representing speech, music, and environmental sounds. By training on a diverse multimedia corpus with layer-to-layer distillation, it achieves near state-of-the-art performance across frame-level speech tasks, audio tagging, and sound classification.

UniVLA: Unified Vision-Language-Action Model

AI Lab: CASIA / BAAI / Tsinghua University / HKISI Summary: UniVLA reformulates vision, language, and robotic actions into shared discrete tokens and learns them jointly in an autoregressive transformer, eliminating separate modality encoders or mapping modules. This unified approach, trained on large-scale video datasets, sets new benchmarks on multi-stage robot manipulation tasks like CALVIN and LIBERO.

ProtoReasoning: Prototypes as the Foundation for Generalizable Reasoning in LLMs

AI Lab: ByteDance Seed / Shanghai Jiao Tong University Summary: ProtoReasoning introduces “reasoning prototypes”—abstract Prolog and PDDL templates—that capture common logical patterns across diverse tasks and guides LLMs to translate problems into these prototypes. Automated prototype construction and verification via interpreters boosts model generalization and reasoning performance on out-of-distribution benchmarks.

Reinforcement Learning Teachers of Test-Time Scaling

AI Lab: Sakana AI Summary: This work trains compact “Reinforcement-Learned Teachers” that ingest both questions and ground-truth solutions to learn dense rewards aligned with student performance, departing from sparse-reward paradigms. A 7B-parameter teacher model surpasses much larger reasoning models on competition-level math and science benchmarks and transfers zero-shot to novel tasks.

🤖 AI Tech Releases

Gemma 3n

Google released a full version of Gemma 3n, its mobile optimized model.

Gemini CLI

Google open sourced Gemini CLI, a coding terminal agent powered by Gemini.

Manus Browser

Manus released an agentic browser.

Qwen-VLo

Alibaba open sourced Qwen-VLo, an image understanding and generation model.

🛠 AI in Production

Project Vend

Anthropic showcased Project Vend, a system that allows Claude to run a small shop.

Ray at Pinterest

Pinterest shares how they scale end-to-end ML pipelines with Ray.

📡AI Radar

Meta has successfully recruited Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai—founders of OpenAI’s Zurich office—to its new “superintelligence team” in what’s being called Zuckerberg’s latest recruiting victory.

Anthropic launched its Economic Futures Program to support research and policy development.

Uber is in talks with Travis Kalanick to find autonomous car company Pony.AI.

Prediction market Kalshi closed a $185 million Series B round led by Paradigm at a $2 billion post-money valuation, even as rival Polymarket reportedly eyes a $200 million raise.

Data management firm Rubrik announced an agreement to acquire Predibase to speed enterprise adoption of agentic AI—from pilot deployments to production at scale.

Battery startup Nascent Materials emerged from stealth after raising $2.3 million to commercialize an energy-efficient process that produces uniformly sized LFP cathode particles for higher-density, lower-cost batteries.

E-commerce veteran Julie Bornstein’s startup Daydream is launching an AI-powered chatbot tailored for fashion shopping following its $50 million seed round.

AI medical scribe Abridge secured $300 million in a Series E to double its valuation to $5.3 billion, led by Andreessen Horowitz with participation from Khosla Ventures.

Voice-to-text app Wispr Flow raised $30 million in Series A funding from Menlo Ventures (with NEA, 8VC, and angel investors) to scale its AI-powered dictation software across Mac, Windows, and iOS.

Andy Konwinski, co-founder of Databricks and Perplexity, pledged $100 million of his own funds via the Laude Institute to back AI research grants and the new AI Systems Lab at UC Berkeley.

Legal-focused AI startup Harvey AI raised $300 million in Series E funding at a $5 billion valuation—just four months after its prior $3 billion round—to expand its automation tools beyond law into professional services.

European challenger bank Finom closed a €115 million Series C led by AVP, bringing its total funding to ~$346 million as it ramps up AI-enabled accounting and targets 1 million SMB customers by 2026.

Creatio unveiled its 8.3 “Twin” release, embedding a unified conversational interface and new role-based AI agents for CRM and workflow automation along with AI-powered no-code development tools at no extra cost.

Nvidia shares have surged back to a record week, positioning the company within striking distance of a $4 trillion market capitalization as demand for its AI chips continues to accelerate.

Audos, the AI-powered startup studio, aims to democratize entrepreneurship by using AI agents and social-media distribution to launch up to 100,000 companies annually without taking equity.

[crocrubies]

stitch by stitch

[marokra]

something about movieverse Sage interests me. i’ve seen a lot of concepts, theories, and ideas thrown around and i adore every single one of them, but honestly i have to wonder why Sage would be created in the first place.

Both she and Stone are both driven by the same thing—loyalty, the only difference being that the former’s coding had that as it’s basis. fundementally, at least from movie 1 Robotnik’s point of view, they serve the same purpose, to protect him, to serve his whims and carry out orders to a tee. having two while only one worked perfectly fine would be redundant, again, from his pov, therefore there wouldn’t be any reason to pursue Sage’s creation. well, unless there was some sort of need.

maybe she was created to assist Robotnik on that mushroom planet, or as a post-sonic 3 thing with fix-it fic undertones.

maybe she was a years-old passion project, some scrapped lines of code he never had the time or purpose to pursue, as she wasn’t particularly needed. he didn’t need a hyperintelligent ai that was built purely to protect and aid him, as Stone did that job well enough already, despite being oh-so-painfully human. so that leads me to wonder which circumstances would drive Robotnik to pursue this dead end, to finish what he started.

there’s a lot of possibilities that could lead to it, honestly. mainly driven from the idea of separation, at least how i see it.

maybe he based her personality on Stone, just a little, most likely unintentionally. deriving from his loyalty, maybe a stray mannerism here and there. Sage, once sentient, once she gets introduced to him, i feel like she’d start to notice the little similarities within her code.

not much gets past an AI, really. she noticed the agent’s quirks, and upon doing a deep dive of her own code, she’d come to realize she had ended up adopting those same mannerisms, that unwavering loyalty towards her father, despite not having known the agent long enough for the mirroring to kick in. it intrigues her. what about the man would drive her father to allow her to mimic him? to deem those traits important enough to include in her code?

but as she kept observing, cataloguing even the simplest of things; like the way he made lattes, his thinly veiled distaste for humanity, and the way he looked at her father like he was the embodiment of the scorching, sharp, yet ever so radiant sun, was when the pieces started to fall into place.

noticing the things that her father loved about his assistant (even though he would deny it to hell and back if she brought up her hypothesis) answered her questions quite clearly.

she knew regular children take on the image of both of their parents. and if her theory was correct, maybe she would come to see Agent Stone as her father, too.

[yaoiwars]

YAOI WARS COMIC #1: LEON GOES TO WORK

[zanukavat]

gonna be a bit mentally ill here for a second

[chancesdealt]

How are you handling everything? Have any of you found 7n7 yet?

🎲 ; " Me? Ever since I got out of that freakish house I've been a lot better, thanks for askin' pal! "

🎲 ; " Yeah, yeah, I was considering going out to find 007n7 soon, miss that guys company. Though most of the survivors gave me the short end of the stick when it came to asking them. So I guess I'll head out alo- "

. . . ?

🔨 ; " Mind an acquaintance ? "

now you can ask both chance and Builderman . Woah!! Yay!!

[jakehal-sometimes]

Imagine jakehal. But yuri. That is all. :3

4

only if it's toxic doomed android yuri. yay 💛