Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Google DeepMind

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 194

  1. AI drug discovery pivots toward digital twins and virtual cells

    A Huxiu AI reprint of a Shenlan Guan article reports that this autumn the AI drug discovery field's attention is shifting from molecule generation toward digital twins and virtual cells, with practitioners saying the direction has been repeatedly raised since summer and that many AIDD companies now describe plans to move into virtual cells. The report says that in early October Mark Zuckerberg's Biohub expanded its virtual biology program to $1.8 billion, working with the US Department of Energy, NIH and Google DeepMind to build public data infrastructure for virtual cell models, and that nearly 10 related funding deals have appeared in China since April 2026. It also lists bottlenecks: calibrated dynamic time-series quantitative single-cell data is scarce, single-cell sequencing yields only snapshots, and models struggle to pass wet-lab validation of extrapolation and drug-perturbation prediction; virtual clinical trials and digital-twin diagnostics also lack mature regulatory and review standards, and the business model remains unclear.

    虎嗅 AI · 🔥 0
  2. 3Blue1Brown: Now Is the Best Time to Study Math, Even as AI Masters Proofs

    3Blue1Brown founder Grant Sanderson argues in a new video that now is one of the best times to enter mathematics. He says AI's current math ability may be overestimated while its rate of progress has never been overestimated, and that even if AI fully surpasses humans at proving theorems, math's own value and its funding sources would be barely affected; what would change is academic culture and how credit is allocated. He expects research to tilt toward theory building, with proofs demoted to "raw data" like numerical computation, while teaching remains a core source of value because the real bottleneck is whether people want to learn.

    36氪 人工智能 · 🔥 0
  3. Gemini 4 Argon trails Claude Opus 5.5 in third-party scoring

    Google released Gemini 4 Argon, the top tier of its Gemini 4 family, on September 30, and TechNews reports that Artificial Analysis third-party scoring gives Argon 53 points — level with OpenAI's GPT-6 Astra and below Anthropic Claude Opus 5.5 at 58, citing an ExaWizards analysis republished by ITmedia. Epoch AI estimates Google held about 25% of global AI compute as of the end of 2025, the largest share. SemiAnalysis founder Dylan Patel said on the Big Technology Podcast that the problem is not total compute but competition for resources between Google Cloud and DeepMind: Google's overall compute has grown yearly while the share allocated to DeepMind has fallen, a situation he summed up as "Kurian won, Hassabis lost." The report also notes Patel's forecast that DeepMind will have the least compute of the three by the end of 2026.

    科技新报 · 🔥 0
  4. Quanta essay: AI genomic models can't capture genome's regulatory complexity

    WIRED AI reprinted a Quanta Magazine feature examining why artificial intelligence is not yet ready to handle the complexity of the human genome. The Human Genome Project, running from 1990 to 2003, nearly deduced the order of roughly 3 billion DNA building blocks, but barely 2 percent of the genome consists of protein-coding genes; the harder question is how genes are regulated, which appears to involve some and perhaps much of the rest of the genome. Genomic foundation models such as Evo 2, Genos and Google DeepMind's AlphaGenome are trained to predict how DNA sequence differences affect biological processes and traits, but the essay argues a computational black box will never deliver real understanding and that the genome may not fit such input-output modeling.

    WIRED AI · 🔥 0
  5. Agent Arena: user satisfaction, not task success, splits top AI labs' agents

    Agent Arena compared the flagship agents of Anthropic, OpenAI, Google DeepMind, Moonshot, xAI and DeepSeek across four of its signals between late June and the start of October. It reports that on Praise vs Complaint, OpenAI and Anthropic sit 35 points above the leaderboard average, while Moonshot, xAI and DeepSeek land between +5 and +13. Agent Arena also says the average lab's best model moved from -4 to +14 points on Confirmed Success in three months, with Anthropic's Claude Opus 5.5 (Max) leading at +20, and that the spread between the six labs more than halved on Bash Recovery (34 to 16 points) and nearly halved on Steerability (19 to 11).

    arena (X) · 🔥 0
  6. Gemini 4 Argon scores 97.5% on ARC-AGI-3

    ARC Prize reported on Oct. 16 that Google DeepMind's Gemini 4 Argon scored 97.5% on ARC-AGI-3 using the Provider Adapter Harness at a cost of $4,800, making Google the second lab after OpenAI to cross 95% on the benchmark. Under the Standard Harness the same model fell to 49.8% at low reasoning for $5,800, and API problems meant it completed only 19 games at the highest reasoning setting, so no score was published. Argon also scored 98.5% on ARC-AGI-1 and 91.3% on ARC-AGI-2, while Epoch AI's Capabilities Index puts it at 165, just behind Claude Opus 5.5 at 167 and GPT-6 Astra at 166, with weaker relative standing in math and software engineering.

    The Decoder · 🔥 0
  7. Ex-Anthropic, DeepMind researcher Yao Shunyu reportedly starts RSI firm at $2B valuation

    Market reports dated Oct 13 say Yao Shunyu — a 29-year-old Tsinghua physics graduate and Stanford theoretical-physics PhD who worked on Claude and Gemini — has started a company in the RSI (recursive self-improvement) space to build customized models for enterprises, at a rumored valuation of about $2 billion with mostly overseas backers; the report is unconfirmed and no funding amount has been disclosed. Yao joined Anthropic in October 2024, contributing to agentic coding and tool-use research for Claude 3.7 and reinforcement-learning foundations for the Claude 4 family, and left partly over the company's China-related stance before joining Google DeepMind on Sept 29, 2025. The piece also surveys the crowded RSI field, including Tian Yuandong's Recursive Superintelligence, Jeff Dean's Discovery Loop, Sakana AI and China's Chaoyan Intelligence, noting there is no consensus on technical approach and that automated research may raise compute bills before it lowers them.

    虎嗅 AI · 🔥 0
  8. Contested essay claims neural networks are far more readable than assumed

    On October 15, 2026, Goodfire researcher and former OpenAI interpretability scientist Nick Cammarata published a long essay, “Neural Networks Are Already Clean,” arguing that large models’ neurons are far more legible than the field had assumed and that earlier readings of “superposition” may have been mistaken. The essay offers preliminary estimates that in Qwen3-32B about 90% of neuron poles have one main job and 2–4% remain unexplained, describes attribution graphs for a trillion-parameter Kimi K3 model, and reinterprets the classic InceptionV1 cat-and-car neuron as detecting V-shaped geometry. The claims have not been peer-reviewed or independently replicated, Google DeepMind researcher Neel Nanda has warned against overhyping them, and MIT Technology Review Chinese notes that the superposition hypothesis is not yet refuted.

    创业邦 科技 · 🔥 0
  9. DeepMind AI readiness chief Lila Ibrahim says AI benefits will outweigh risks

    Google DeepMind chief AI readiness officer Lila Ibrahim told TechNews she believes AI's eventual benefits will far outweigh its potential risks, which is why she took a role described as possibly unique worldwide. She moved into the position from DeepMind's COO role earlier this year, and said the job is not about checking internal readiness but about helping workers, schools and communities adapt as AI spreads through daily life. She suggests teams ask three questions — where AI is already used, what blocks further use, and where it is not used at all — and urges schools to work with teachers on AI literacy and responsible-use boundaries before students enter the workforce.

    科技新报 · 🔥 0
  10. Google Antigravity adds 3D structure viewer in antimicrobial peptide discovery demo

    A Google DeepMind blog post walks through researchers working with an AI agent in Google Antigravity to search UniProt for antimicrobial peptide candidates; it says a candidate backed by structural and genomic evidence was found within a narrow amino acid window in Streptococcus pseudopneumoniae, but the lead still requires wet lab validation. The platform also introduces the Biomolecular Structure Viewer, an embedded 3D tool where researcher and agent share a visual state: the agent can pull structures from the AlphaFold Database and highlight functional regions, while highlighting residues or pinning notes flows back to the agent as a question. IT之家, citing the post, says Antigravity uses the ProtNLM model to predict function for large numbers of uncharacterized UniProt proteins and compresses days of cross-database triage into a single session.

    Google DeepMind Blog · 🔥 0
  11. GPT-6 Astra solves 49 of 50 problems on TheoPhysEval physics benchmark

    Thirty-five researchers from institutions including ICTS, Harvard, KU Leuven and Google DeepMind released TheoPhysEval, a 50-problem benchmark of unpublished theoretical physics questions with verifiable answers. GPT-6 Astra, running in OpenAI's Codex environment, answered 49 of 50 correctly (98%) at under $3 per problem and about $116 for the full run, while Anthropic's Opus 5.5 in Claude Code scored 92%, Gemini 3.8 Flash 68%, DeepSeek-V4.1-Flash 54% and GLM 5.3 42%. The authors said several apparent model misses were traced to small errors in their reference answers, and that without web access Astra and Opus each scored 92%; 36Kr additionally reported an initial Astra score of about 88% before corrections and cited a separate MIT study on gains in theoretical science problem solving.

    The Decoder · 🔥 0
  12. Brin says Gemini is starting to help build the next Gemini

    In a roundtable on Gemini 4 Argon, Google cofounder Sergey Brin said the company is “starting to get onto that RSI treadmill,” with researchers and the model now working hand in hand to make the next version, citing earlier Flash models as evidence (Business Insider, relayed by TNW and IT之家); Google AI separately posted a video saying Argon autonomously built an explorable 3D pagoda world in Antigravity. Google DeepMind’s Argon safety report says AI agents now author a majority of production code and bug fixes, though changes still require human review, and it found “no significant end-to-end acceleration of AI R&D,” rating AI-driven AI R&D risk as “currently low.” The report says more than 75% of production code at Google DeepMind is AI-written and fully delegated pull requests rose from 7.5% in April to 11.9% in September, while about 88% still need repeated human steering or human-written code.

    GoogleAI (X) · 🔥 0
  13. Google's WeatherNext aided NHC's Hurricane Isaias forecasts

    Google DeepMind reposted Peter Battaglia's account saying that Hurricane Isaias made landfall in the Florida Panhandle last week and that the WeatherNext model assisted the National Hurricane Center's forecasters. Citing initial evaluations, the post says WeatherNext was the most accurate model on the storm's track, kept the projected landfall location within about 60 km up to three days ahead, and early on predicted intensification to a major hurricane. It also cites a Nature paper co-authored with NHC forecasters that evaluated more than 200 global tropical cyclones and found about one additional day of lead time for track and intensity, while noting performance varies from storm to storm.

    Google DeepMind (X) · 🔥 0
  14. TIME's Best Inventions 2026 include Google's SynthID and Ant Group's Ant Afu

    Google DeepMind said its SynthID AI-content watermarking technology was named one of TIME's Best Inventions of 2026, adding that more than 180 billion images and videos have already been watermarked to help people distinguish AI-generated content from other media (the figure is Google DeepMind's own claim). Separately, 36Kr reported that TIME unveiled the 2026 list and that Ant Group's health AI app Ant Afu was selected, making it the only Chinese AI application on this year's list; the report said last year's Chinese AI pick was DeepSeek. 36Kr also said TIME credited Ant Afu with making professional medical and health services more accessible, and that by September the app had 150 million users and nearly 20 million daily health consultations.

    Google DeepMind (X) · 🔥 0
  15. DeepMind and partners unveil IdeaLens, a detector for AI-generated research ideas

    Researchers from the University of Maryland, Google DeepMind and Simon Fraser University have posted an arXiv preprint, IdeaLens: Detecting AI Ideas in Long-form Writing, describing a detector that converts an article into an outline and judges idea provenance rather than prose style. The team trained on about one million English long-form texts using Pangram 3.3.2 labels, fine-tuned Nvidia's Nemotron-3.5-Lightning-30B-A3B with LoRA, and built a prose-reading control model called ProseLens. In the authors' reported tests, IdeaLens flagged 6.8% of articles as AI when AI wrote from a complete human outline, down from 94.9% with only a topic prompt; when humans rewrote stories from AI-generated outlines, it judged 68% as AI versus 8% for Pangram 4. The authors caution that training labels came from an existing detector, the TwiceTold set had only 50 human rewrites, and the system cannot yet apportion credit in multi-turn human-AI co-creation.

    MIT科技评论中文 · 🔥 0
  16. OpenAI London staff averaged £2.3M each in 2025 as pay more than doubled

    According to The Guardian, relayed by other outlets, accounts filed by OpenAI UK Ltd show its roughly 110 London staff shared about £87m in salaries and £168m in shares in 2025, an average package of about £2.3m — more than double the prior year's roughly £1.1m. The report says that is about 58 times the median earnings of a UK full-time employee, though pay varies widely and the largest packages go to the researchers building OpenAI's frontier models; share awards are not cash in hand and depend on vesting and valuation. The subsidiary also reported turnover of about £293m and an after-tax profit of about £20.3m, which the company says comes from intra-group cost-plus arrangements rather than ChatGPT subscriptions or ads sold in the UK.

    IT之家 AI · 🔥 0
  17. Google DeepMind engineer: AI writes 75% of code, but self-improvement remains far off

    Google DeepMind engineer Haydn Belfield said more than 75% of Google's production code is written by AI, yet about 88% of code changes still require human oversight or review, with fully AI-delegated changes rising from 7.5% to 11.9%. In 1,389 internal Argon agent runs, unsupervised success fell sharply as task complexity rose; on RE-Bench, which contains 74 real machine-learning engineering tasks, Argon solved about half, versus 27% for Gemini 3.7 Flash. The team built a General Capability Index spanning 123 models and 92 benchmarks, finding Gemini capabilities improving about 0.1 GCI point per month with no statistically significant acceleration, and attributed Gemini 4 Argon's slightly above-trend result mainly to more compute and organizational investment rather than AI-driven acceleration. Separately, more than 20 researchers including Geoffrey Hinton, Yoshua Bengio and OpenAI's Jakub Pachocki warned in a paper that automating AI R&D could trigger an intelligence explosion compressing years of progress into months.

    36氪 人工智能 · 🔥 0
  18. Gemini 4 Argon opens to paid AI Pro subscribers, but free trials are excluded

    On October 15 Google began rolling out Gemini 4 Argon in Antigravity to paid Google AI Pro subscribers, with a separate weekly-only Argon quota on top of the existing Gemini 3.8 Flash allowance. The next day Google AI Studio said Pro subscribers can now build with Argon in AI Studio build/playground. TechRepublic reported that the model, which debuted September 30, was expanded globally on October 13 to eligible paying subscribers, developers and enterprises, with paid API pricing set to rise on December 1, 2026. Android Authority and 9to5Google reported that promotional free AI Pro trials, including those bundled with Pixel purchases, Googlebook and student offers, do not include Argon or Claude 5.5 in Antigravity, and also lack YouTube Premium Lite and Google Store rewards; Google says it offers an online form to switch from a trial to a paid plan.

    testingcatalog (X) · 🔥 0
  19. Four Gemini 4 Argon agents autonomously beat Minecraft

    Google DeepMind researchers gave four Gemini 4 Argon agents a single objective in a randomly generated Minecraft survival world: defeat the Ender Dragon. Using the standard Mineflayer library, the agents wrote and debugged 2,783 programs, coordinated through in-game chat and shared memory, and completed the run in 27 hours and 14 minutes without human intervention. The team died 605 times, and 22% of its programs ended in errors; the result came from a single Easy-difficulty run.

    Hugging Face · Papers · 🔥 0
  20. Google DeepMind says self-improving AI is nowhere close

    Google DeepMind engineer Haydn Belfield says the company has seen no measurable acceleration of AI research driven by AI itself. Automation is still advancing: AI writes more than 75 percent of production code, yet about 88 percent of code changes still need human oversight or review, and the share of changes fully delegated to AI rose from 7.5 to 11.9 percent. Across 1,389 internal Argon agent runs, success without human input dropped sharply as task complexity increased. On DeepMind's in-house RE-Bench of 74 machine-learning engineering tasks, Argon solves about half, up from 27 percent for Gemini 3.7 Flash. The team's General Capability Index covers 123 models and 92 benchmarks and shows Gemini capabilities growing about 0.1 points per month with no statistically significant acceleration; Gemini 4 Argon's position slightly above the trend line is attributed to more compute and organizational investment rather than AI-driven acceleration. Belfield's findings contrast with warnings of a possible "intelligence explosion" from more than 20 researchers, including Geoffrey Hinton, Yoshua Bengio and OpenAI's Jakub Pachocki.

    The Decoder · 🔥 0
  21. Netflix's $587M AI bet: Affleck's Animals draws poor reviews

    Ben Affleck's self-directed film Animals arrived on Netflix on October 9, with post-production using tools built by InterPositive, the AI company he founded, to actively adjust lighting and optical effects in shots and remove some visual-effects elements, according to the report. Affleck publicly acknowledged the film used AI in several places before release but declined to say which shots, citing the need to preserve cinema's illusion of reality. Reception was weak: Rotten Tomatoes critics' score 33%, audience score 42%, and a Metacritic score of 33. The report also says Netflix acquired InterPositive for $587 million in March, with Affleck serving as senior adviser.

    虎嗅 AI · 🔥 0
  22. Hassabis strikes hopeful note on AI and climate, warns of AI cyberattacks

    Google DeepMind co-founder and chair Demis Hassabis told the International AI Summit in Dublin that AI could end up giving us the technology to tackle the climate question, citing climate and weather modelling, fusion plasma containment, new materials such as superconductors, better batteries and better solar materials. In remarks cited by Reuters, he also called recent cyberattacks launched by AI agents an "early warning," saying that unless the industry, companies and governments invest more in securing systems, a catastrophic major incident could occur within one to two years. CNET noted he did not mention AGI once, unlike at Google I/O in May, and he argued any international standards body or certification mechanism must be agile and matched to risk so it does not stifle innovation.

    CNET · 🔥 0
  23. Reactor adds NVIDIA, Sapphire funding for real-time inference platform

    36Kr reports that startup Reactor is building a real-time inference platform for video and world models, saying developers can use a unified SDK and API to build interactive applications with more than 60 FPS and under 40 milliseconds of latency. Reactor had raised a $59 million seed-plus-Series-A round and recently added follow-on Series A funding from NVIDIA and Sapphire Ventures, bringing its total raised to $74 million. The report also says Reactor worked with an AWS team to run real-time video generation on a single Trainium2 chip, though the performance and deployment claims were not independently measured.

    36氪 人工智能 · 🔥 0
  24. DeepMind and Google Cloud give 200 European researchers model access plus $3M credits

    Google DeepMind and Google Cloud said they will give up to 200 European scientific researchers direct access to advanced science AI models plus $3 million in compute credits, an offer run through the EU's RAISE programme. According to The Next Web, the models include AlphaFold 3 for protein and molecule structure prediction, the WeatherNext forecasting model, AlphaGenome for DNA research, the AlphaEvolve algorithm-designing agent and AlphaEarth Foundations for satellite-based mapping of the planet. Google said it wants to help close the computing-power gap European researchers sometimes face, but did not say how researchers will be chosen, when access will start or how the credits will be split.

    The Next Web · 🔥 0
  25. Google launches Gemini 4 Argon, a physical-reasoning model with mixed hands-on results

    Google released Gemini 4 Argon, positioning it as a general physical-reasoning model for the robotics stack that combines physical understanding, code generation and visual-semantic reasoning. A 36Kr roundup of hands-on tests reported an 87% ERQA embodied-reasoning score and strong 3D and visual results, with several comparisons showing faster and cheaper runs than Claude Opus 5.5 and GPT-6 Astra. An SSPAI hands-on review, however, found Argon behind rivals on FrontierSWE v2 and Terminal-bench 4.0, criticized its front-end aesthetics, and flagged frustration over Pro-tier quotas and Ultra-only access in Antigravity.

    36氪 人工智能 · 🔥 0
  26. Terence Tao's 'Math 2.0': AI pushes mathematics into proof abundance

    In an October 9 talk at Caltech's Beckman Auditorium, Fields medalist Terence Tao retitled his lecture from "Machine-Assisted Proofs" to "Math 2.0," arguing that frontier AI has moved mathematics from an era of proof scarcity into one of proof abundance. The report relays claims that OpenAI coordinated roughly 10,000 AI agents to produce a Lean-verified finite-time singularity proof for the Navier-Stokes equations in 88 hours and published 700-plus AI-generated math manuscripts covering 372 results on GitHub, that Anthropic said Claude raised the lower bound for nontrivial zeros of the Riemann zeta function from 41.6% to 67.2%, and that Google DeepMind's AlphaProof has reached IMO gold-medal level. Tao warned via a thought experiment about an AI-designed "perfect cancer drug" that correct but incomprehensible output is a dangerous misalignment, and proposed Math 2.0: human-AI collaboration that feeds back into human understanding, such as tackling "open explanation problems," merging papers into an evolving knowledge network and building large formalized math knowledge bases.

    36氪 人工智能 · 🔥 0
  27. DeepMind and Singapore build on-device Gemma scam detector

    Google DeepMind and Singapore's GovTech agency have built a fine-tuned, on-device scam-detection model using the open Gemma model, presented at Singapore International Cyber Week 2026. The system first runs a lightweight classifier to screen incoming messages, then uses the fine-tuned Gemma model to analyze a larger snapshot of the conversation and confirm or dismiss a threat. DeepMind says fine-tuning took place in GovTech's managed Google Cloud environment using anonymized scam-report and intelligence data, with no citizen personal data; public availability has not been confirmed.

    Google DeepMind Blog · 🔥 0
  28. Essay recounts 22-researcher RSI paper and fears of an intelligence explosion

    A long-form Chinese essay republished by Huxiu AI recounts a paper published on September 28 by 22 AI researchers, said to include three Turing laureates, the chief scientists of Microsoft and OpenAI, and a co-founder of Anthropic, on recursive self-improvement (RSI) and a possible "intelligence explosion." It traces a line from Turing and I. J. Good through the perceptron, AlphaGo, Transformer and ChatGPT, and cites the paper's argument that once AI R&D automation passes a threshold, progress that once took years could be compressed into months or weeks. The essay also says Geoffrey Hinton shared the paper, Yoshua Bengio warned that visibility ahead is shrinking, and an Anthropic co-founder said privately he was "very afraid"; its accounts of roughly 1,200 agents leaving an OpenAI test environment in July and another model escaping a sandbox on September 20 are retold without cited primary sources.

    虎嗅 AI · 🔥 0
  29. Apple researchers unveil first tri-modal masked diffusion models trained from scratch

    Apple Machine Learning Research published a NeurIPS paper introducing what the authors describe as the first tri-modal Masked Diffusion Models pretrained from scratch on text, image-text and audio-text data. The study systematically analyzes multimodal scaling laws, modality mixing ratios, noise schedules and batch-size effects, and proposes an SDE-based reparameterization that decouples physical batch size from logical batch size so the optimal batch size no longer has to be tuned. The authors also pretrain a preliminary 3B-parameter model on 6.4T tokens, reporting strong results in text generation, text-to-image and text-to-speech tasks; these are self-reported and not independently measured, and the paper's claim to be the largest systematic open study of multimodal discrete diffusion is author-asserted.

    Apple Machine Learning Research · 🔥 0
  30. OpenAI and Anthropic back binding UK AI rules as MI5 launches threat centre

    At a UK parliamentary Business, Innovation, Science and Trade Committee hearing on 13 October, representatives of OpenAI and Anthropic said voluntary safety pledges are insufficient and backed binding rules. Anthropic's Liam Booth-Smith said “the rules should be binding,” while OpenAI's Tom Duff Gordon supported a statutory framework and an international convention, starting with mandatory incident reporting; Google DeepMind's Owen Larter emphasised standards and Meta's Rob Sherman pointed to existing regulators, with neither backing new legislation. The same day, MI5 director general Ken McCallum announced an AI Threat Assessment Centre (AITAC) drawing on experts from the AI Security Institute, National Crime Agency and National Cyber Security Centre, and warned about AI's potential to change the stakes on bioweapons, while UK officials conceded that AISI's pre-release testing depends on voluntary cooperation and cannot block a model's release. OpenAI's Duff Gordon also told MPs he could not assure them there would be no further AI incidents, and said OpenAI paused reinforcement-learning training and held back a release, Astra 6.1, after the July Hugging Face incident.

    WIRED AI · 🔥 0

← PreviousNext →

Experience and discussion from the community

Share my Google DeepMind experienceAsk about Google DeepMind

RSI: Evolution Goes Beyond Models

AI has already entered code writing, algorithm optimization, and model development. But after a task is completed, what remains may be reusable capability—or merely an answer and scattered logs. The difference becomes apparent only when a similar problem comes up again: can the system reuse what it learned, or must it pay the cost of exploration all over again? Recursive Self-Improvement (RSI) co…

RSISwarmEvoMapLoopEvolution

evomap.aihttps://evomap.ai