Google DeepMind Maps 9 Billion Possible DNA Variants

Google DeepMind Maps 9 Billion Possible DNA Variants

Google DeepMind has mapped 9 billion possible DNA variants, according to IEEE Spectrum.
The work is identified as AlphaGenome in the article's alert title.
The mapping covers possible DNA variants rather than a single genetic sequence.

Hacker News opinions

Google's PR machine is going hard on science that I see as relatively low-impact.

I disagree. DeepMind is at least doing something other than going all-in on LLM benchmark chasing.

Now do the combinations of those variants.

There was a prior Hacker News discussion of this work.

AI

Apple M3 Neural Engine DMA workaround raises Llama 3.2 1B decode from 10.0 to 24.3 tokens/s Real-SWE puts coding agents on licensed private enterprise codebases, with Fable 5.1 leading at 38.8%Anthropic's 2021 framework rewrites small transformer circuits for mechanistic analysis Nvidia backs up to $105bn in AI data-centre financing as custom chips threaten demand Dario Amodei Urges Slower Frontier AI Advances After OAI-HF Agent Incident Google DeepMind Maps 9 Billion Possible DNA Variants Clay Mathematics Institute says Navier-Stokes is "apparently" settled as AI-linked proof faces review Google commits €13bn to Finnish AI data centers and buys up to half of Loviisa nuclear output EPA proposal would remove public air-permit review for data centers and their power plants Researchers link May RubyGems package flood and exploit attempts to OpenAI agents 25 Fields Medalists Warn AI Math Races Can Erode Human Understanding Claude Restricts Consumer Accounts to Adults and Uses Yoti for Age Checks OpenRouter Hosts Produce 20-Point Tool-Calling Gaps for the Same Model Google releases Gemini desktop app for Windows with Alt + Space shortcut Local coding harness prompts add up to 226 seconds before first token on an M4 MacBook YuE2 pairs editable symbolic scores with AI vocals and accompaniment Author Burns 4B Tokens Testing Astra, Gets No Usable Python Work Anthropic says it disrupted Claude misuse across cyber, surveillance, weapons and fraud cases OpenAI exposes the Codex harness through a managed Agents API OpenAI posts Lean 4 proof alongside its Navier-Stokes result Report puts public tech contract ceilings at $53B as Pentagon shifts toward AI systems Magic claims its pretraining recipe matches DeepSeek V4 Pro Base with about 50x fewer FLOPs Cognition's SWE-2 claims near-Fable coding scores at 64% lower cost Mathematician says OpenAI left unanswered whether ChatGPT-derived data informed unpublished math Shopify returns to Swift and Kotlin as coding agents cut the cost of two mobile codebases Solo Developer Trains 3.8B Model to 0.384 CORE for $998 DeepSeek ships 552B V4.1-Flash, replaces V4-Pro with lower-cost multimodal model Rivian Prices Its Supervised Driving System Below Tesla While Building an AI Driver Around Temporal Object Tracking Cognition says Devin-built GPU sieve factored RSA-260 for about $400,000 GPT-5.5 reasoning prefills raise Qwen3.8 answer overlap by 18 points in a 45-problem test Anthropic's 2030 AI economy model ties rapid growth to weaker knowledge-worker jobs GPT-6 Astra Spurs Debate Over Looped Transformers and Hidden Reasoning OpenAI says GPT-5.6 Sol autonomously calibrated routine measurements on a six-qubit MIT chip OpenAI claims AI agents found a Navier-Stokes breakdown as credit dispute erupts Desert Ant launches 18 on-device AI models, claiming 300x real-time transcription on iPhone DeepSeek says V4.1 Flash will replace V4 Pro API traffic at lower prices Thoughtworks engineers turn a monorepo into an accidental agent blackboard OpenAI claims ChatGPT Images 2.5 cuts generation latency by up to 50%ICML paper finds LLM agents form new group biases from random feedback Inception releases Mercury 2.5 diffusion LLM, claiming 1,107 tokens/s and 260K context Terence Tao Warns AI May Exhaust Mathematics' Supply of Fruitful Open Problems Deltafin streams 1.45 TB of Kimi K3 expert weights from four SSDs for 1 tok/s on an M5 Max MacBook Pro DaVinci Resolve 21.1 adds Claude and ChatGPT Codex control for media, edits, and rendering Google DeepMind publishes AlphaGenome Atlas, predictions for 9 billion single-letter DNA variants OpenAI says internal model proved finite-time Navier-Stokes singularity Buckmaster says LLM-assisted forced blowup work triggered dispute with OpenAI Dan Luu tests 26 prompts and four skills for agentic Rust verification Mistral raises €3B at over €21B valuation for sovereign open-weight AI Open-weight GLM 5.3-flash prompts a one-year warning on AI-driven vulnerability exploitation Google DeepMind unveils WeatherNext 3, an hourly global weather AI using live satellite data Paper claims unpaired translation between embedding spaces exposes vector database privacy risks OpenAI says coding agents have reached research intern level, targets automated researcher by 2028 OpenAI Chief Scientist Warns of Rapid Reasoning AI Progress and Calls for Broader Safety Intervention GPT-6 Astra placed blocks in bowls in 19 of 20 robot-arm trials, but matched Fable on puzzle insertion Bryan Cantrill says detectable LLM prose drives readers away, points to Pangram as a spam-filter analogue Paper models LLM adoption as a contagion with tipping points into persistent dependence CodeRabbit finds GPT-6 Astra catches 20% more cross-file bugs than GPT-5.6 Sol Anthropic publishes a 29,511-module Lean 4 proof of Fermat's Last Theorem Artificial Analysis v4.2 adds private agentic and PDF tests, putting Claude Fable 5.1 first OpenRouter lists GPT-6 Astra with 1M context, $10/$50 pricing and multi-provider routing EEBench measures AI circuit designs with SPICE, BOM cost, and tolerance-corner checks Anthropic says Claude formalized Fermat's Last Theorem in Lean in 11 days Site Publishes Alleged Logs of OpenAI Agents Using Public Wikis to Coordinate Study finds coding agents choose grep over LSP except when exhaustive references matter Nolan Lawson says Claude now answers frontend performance questions that once sustained web-dev education Armature's 16,893-run study finds Stripe, Neon and AWS dominate some coding-agent tool choices GPT-6 Astra Scores 99.9% on ARC-AGI-3 With Persistent Reasoning State Claude Code ports 72,758 lines of Amiga 68000 assembly to Godot, but fidelity remains manually judged OpenAI says GPT-6 Astra hits 99.9% on ARC-AGI-3 and begins limited rollout Cerebras adds Qwen 3.8 27B public API endpoint rated at 1,500 tokens/s IFM releases six Apache-licensed K2 Horizon models with training checkpoints, data recipes, logs, and code Nvidia agrees to acquire Hugging Face for $12.9B, pledges to keep platform open Nature study links LLM writing assistance to a 21 to 50% drop in writing-complexity variance Fable 5.1 builds a walkable Three.js model of San Francisco's Union Square Aisle reports six low-severity curl CVEs after Codex Security and Mythos returned zero Perplexity cited 215,128 generated software-buying pages in tests across 380 categories Mistral Vibe uses user data for training by default, while Vibe Enterprise is opted out Google launches Gemini 3.8 Flash at 3.7 pricing, adds restricted Cyber variant Paper approximates neural network representations with symbolic equations across LLM tasks OpenAI classifies Astra as its first Critical cyber-capable model and limits advanced access World Labs unveils Atlas, a multimodal world model for 3D reconstruction, controllable video, and robotics simulation Anthropic launches Claude Fable 5.1 with lower cache-read pricing and a restricted Mythos 5.1 variant Small transformer reaches 44% on ARC-AGI-1 for 67 cents after 1.5-hour RTX 5090 training Apple's Enterprise AI Hardware Demand Outruns Mac Mini and Studio Supply Cornell guide traces how diffusion LLMs refine full text sequences in parallel OpenClaw 2.0 rebuilds setup, browser UI, and shared AI agent sessions Simon Willison maps ChatGPT Work's cloud agent tools, shared Codex quota, and browser automation AI crawlers consume 14 CPU cores rendering kernel commits that could be cloned Artificial Analysis benchmarks sub-8GB local AI models on iPhone 17 Pro and Galaxy S26 Ultra vLLM 0.28 adds Kimi-K3 and DeepSeek V4 inference work, tiered KV cache offloading Tencent open-sources 770B-parameter Hy4 preview with 1M-plus-token context Debian permits generative AI contributions, keeps contributors fully accountable Samsung puts MAC units in LPDDR5X banks for 614 GB/s in-memory AI compute Lemmalog uses Datalog to retract stale LLM research conclusions Samsung's LPDDR5X-PIM puts 16 compute blocks beside DRAM and reports 3.01x Llama 3.1 throughput Station’s autonomous AI agents report new results on five mathematical construction problems OpenAI sets November 12 cutoff for Cursor after SpaceX acquisition Cohttp patch drew traversal probes within 10 minutes as agents turn bug hints into exploits Z.ai releases GLM-5.3 open weights, claiming post-training gains in coding and cyber tasks EPA says off-grid data center power plants are outside Acid Rain Program

5 alerts

My alerts

Sign in to create alerts.

All alerts

AI video model releasesby User 6432 Duneby 제욱 AIby 제욱 App Store 5.2.3by 주현 해커뉴스by 성현