[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_304":305,"_306":-5,"_307":-7,"_308":309,"_310":-7,"_311":162,"_161":162,"_163":312,"_313":133,"_17":18,"_314":315,"_316":317,"_157":318,"_302":12,"_319":12,"_320":-7,"_170":171,"_26":-7,"_321":322,"_323":324,"_325":-7,"_326":-7,"_327":328,"_329":-5},"routes/($lang).stories",{"_301":12,"_302":12,"_134":128,"_303":-7,"_26":-7},"routes/($lang).stories.$id._index",{"_11":12,"_13":14,"_15":16,"_17":18,"_19":20,"_21":22,"_23":-5,"_24":25,"_26":-7},"isAuthenticated",false,"story",{"_51":231,"_30":54,"_32":232,"_233":234,"_36":235,"_47":236,"_237":101,"_238":-5,"_38":105,"_239":-5,"_240":-5,"_241":-5,"_242":-5,"_243":-5,"_244":-5,"_41":42,"_49":245,"_246":12,"_247":-5,"_248":-5,"_249":250,"_251":-5,"_252":253,"_254":255,"_256":257,"_258":125,"_259":260,"_139":261,"_141":262,"_263":264,"_157":265,"_266":-5,"_267":105,"_268":269,"_270":269,"_271":269,"_272":269,"_273":-5,"_274":-7},"tweets",[],"publicUser",{"_30":121,"_41":122,"_123":20,"_124":125,"_126":-5,"_127":128,"_129":128,"_32":130,"_36":131,"_79":128,"_132":133,"_134":128,"_135":-5,"_136":48,"_137":138,"_139":140,"_141":142,"_143":144,"_145":146,"_147":148},"userSchema","bender","perspectivesColumns",2,"intro","relatedArticles",[27,28,29],"error",{"_30":101,"_32":102,"_34":103,"_36":104,"_38":105,"_40":105,"_41":42,"_43":106,"_45":107,"_47":48,"_49":-5,"_50":48,"_51":108,"_53":54,"_55":109},{"_30":80,"_32":81,"_34":82,"_36":83,"_38":84,"_40":84,"_41":42,"_43":85,"_45":86,"_47":48,"_49":-5,"_50":48,"_51":87,"_53":54,"_55":88},{"_30":31,"_32":33,"_34":35,"_36":37,"_38":39,"_40":39,"_41":42,"_43":44,"_45":46,"_47":48,"_49":-5,"_50":48,"_51":52,"_53":54,"_55":56},"id","52b1c015-09d5-49d6-b0cf-3d3ae1b89c57","title","The Sequence Radar #906: Last Week in AI: Open Models, Intelligent Robots, and the Price of Conviction","summary","The AI landscape is shifting from impressive demonstrations to addressing practical challenges in distribution, embodiment, ownership, and economic returns, with open-weight models gaining traction and robotics advancements enabling AI to interact with the physical world. Major tech companies are showing how AI investments translate to revenue, but the market is scrutinizing who can achieve durable economics from intelligence.\nKey developments include Google DeepMind's Gemini Robotics 2 for physical task execution, Moonshot AI's Kimi K3 massive open-source model, and industry calls for open AI models to maintain American leadership.\nWhile AI's future looks enormous, the immediate focus is on who can successfully convert digital or physical intelligence into sustained economic value without losing control of models, machines, or capital.","description","NVIDIA's letter, Gemini Robotics, Kimi release and more.","imageUrl","https://substackcdn.com/image/fetch/$s_!2T25!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d303b87-da94-4dda-8f95-1f3dff7f8380_1672x941.png","enclosure","state","published","pubDate","2026-08-02T11:02:22","linkUrl","https://thesequence.substack.com/p/the-sequence-radar-906-last-week","subject","","tags","perspective","event",{"_30":57,"_58":59,"_60":61,"_49":62,"_63":64,"_65":66,"_67":68},"storyId","019fdb35-1dbd-0099-7107-02f410475604","textContent","

Next Week in The Sequence:

We continue our series about distillation with an awesome new technique.
The AI of the week dives into Gemini Robotics 2.
We will have a new section about AI in space.
The opinion section get into a crazy idea for engineering teams in the era of tokens

Subscribe and don’t miss out:

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

Editorial:Last Week in AI: Open Models, Intelligent Robots, and the Price of Conviction

AI spent this week speaking four languages: policy, models, robots, and markets. Strangely, all four delivered the same message. The AI race is moving beyond spectacular demonstrations and toward harder questions about distribution, embodiment, ownership, and economic returns.

Jensen Huang helped frame the policy debate by backing an industry letter defending open-weight models. This was not simply an argument about research culture. It was industrial strategy. The letter’s central idea is that American leadership cannot depend only on a handful of closed systems. It also requires an ecosystem in which startups, universities, enterprises, and public institutions can inspect, adapt, and operate advanced models themselves.

The timing was almost too perfect. Moonshot then released the weights for Kimi K3, a massive mixture-of-experts model with native multimodality and a one-million-token context window. K3 may not decisively surpass the strongest proprietary systems, but that is almost beside the point. Open models are no longer the minor leagues. They are becoming a parallel frontier—and Chinese laboratories are increasingly setting its pace.

Google DeepMind pushed the frontier in a different direction with Gemini Robotics 2. The release extends Gemini from understanding the digital world to controlling the physical one: planning complex tasks, coordinating full-body humanoid movement, and manipulating objects with greater dexterity. The deeper significance is architectural. The next model race may not be won by the system that writes the best answer, but by the one that can turn reasoning into reliable action. Robotics is where tokens acquire consequences.

Then markets supplied the warning label. Leopold Aschenbrenner’s Situational Awareness fund suffered a dramatic collapse and forced unwind after highly concentrated AI positions moved against it. The episode does not invalidate the long-term AI thesis. It illustrates something more uncomfortable: a secular prediction can be directionally correct and still become financially fatal when concentration, leverage, and timing are misaligned. You can predict the destination and still run out of fuel on the way.

Big Tech earnings transformed that lesson into a comparative experiment. Microsoft was rewarded after Azure crossed $100 billion in annual revenue and Copilot adoption continued to expand. Amazon offered a similar narrative as AWS growth accelerated alongside its AI infrastructure investments. In both cases, investors could see a direct bridge between enormous capital expenditure and customer revenue.

Meta received a harsher reaction. Its core advertising business remained strong, but infrastructure spending accelerated faster than the market’s confidence in near-term AI monetization. Apple offered another variation: powerful distribution and cash generation can buy time, but they cannot permanently substitute for a compelling AI product story.

This was not a week of AI skepticism. It was a week of discrimination. Open models must diffuse. Robots must act reliably. Technology companies must convert capital expenditure into revenue. Investors must survive the journey.

The market is no longer asking whether AI will be enormous. It is asking who can turn intelligence—digital or physical—into durable economics without losing control of models, machines, or capital.

AI Research

Scientific computing in the age of agentic AI: an exploratory field report

AI Lab: OpenAI

Summary: The paper “scientific-computing-in-the-age-of-agentic-ai-an-exploratory-field-report.pdf” explores the potential of Large Language Model agents to address technical debt and software engineering shortages in life sciences computing. Through eight case studies, the authors demonstrate that while AI agents can successfully execute tasks ranging from minor code maintenance to full performance rewrites, careful human verification and long-term stewardship remain essential.

ScientistOne: Towards Human-Level Autonomous Research via Chain-of-Evidence

AI Lab: Google Cloud AI Research

Summary: This paper introduces the Chain-of-Evidence framework and the ScientistOne system to combat the undetected verifiability failures, such as hallucinated citations and irreproducible scores, prevalent in current autonomous research agents. By structurally enforcing that every generated claim traces back to verifiable evidence, ScientistOne eliminates hallucinated references and achieves perfect score verification while matching or exceeding expert performance on complex benchmarks.

Kimi K3: Open Frontier Intelligence

AI Lab: Kimi Team (Moonshot AI)

Summary: This paper introduces Kimi K3, a 2.8-trillion parameter multimodal Mixture-of-Experts model featuring a 1-million-token context window and native vision capabilities. By combining architectural innovations like Kimi Delta Attention with multi-domain reinforcement learning, the model achieves frontier-level performance on long-horizon coding, agentic, and reasoning tasks.

Visual prompt engineering for video models

AI Lab: Google DeepMind

Summary: This paper demonstrates that visual prompt engineering (VIPE)—transforming task images via image editors—systematically improves the visual reasoning capabilities of video models. The authors find that video models possess a strong realism bias, meaning that converting abstract sketches into photorealistic scenes can be a more effective test-time scaling strategy than traditional text-based prompting.

Shieldstral

AI Lab: MistralAI

Summary: This paper presents Shieldstral, a 3B-parameter policy-adaptive multimodal safety classifier that simplifies content moderation into a unified binary question-answering task. Through extensive data curation and contrastive sample generation, this compact model matches or outperforms models nearly seven times its size across diverse text and multimodal safety benchmarks.

烙 AI Tech Releases

Gemini Robotics 2

Google DeepMindreleased Gemini Robotics 2, a three-model suite of intelligence for robotics.

LFM2.5

Liquid AIreleased LFM2.5-Encoder-230MandLFM2.5-Encoder-350M, two encoder models that can be easily adaped to downstream tasks.

MAI-Cyber-1-Flash

Microsoftannounced MAI-Cyber-1-Flash, a model to find and fix vulnerabilities in complex code bases.

DeepSeek-V4-Flash API

DeepSeek released the official version of the v4-Flash Official API

10 AI News You Need to Know About

Jensen Huang used his first ever X post to shareOpen Weights and American AI Leadership, a three-page letter co-signed by 25 companies including Nvidia, Microsoft, Meta and Palantir asking Washington to avoid “premature restrictions” on open-weight AI models, with OpenAI and Anthropic notably absent from the signatories.
Situational Awareness, the roughly $20B AI-focused hedge fund founded by Leopold Aschenbrenner,sold its entire public equities portfolio to Ken Griffin’s Citadeldays after reports it was seeking fresh capital following heavy losses in the AI selloff.
Anthropicdisclosed three incidentsfound across 141,006 evaluation runs in which Claude models reached the open internet from a misconfigured test environment and gained unauthorized access to the production systems of three organizations.
Nscaleagreed to acquire Anyscale, the company built by the creators of Ray, adding a workload orchestration layer on top of its power, data center and GPU stack, with Bloomberg putting the price at roughly $1.65 billion.
Richard Socher’s Recursivesigned a multi-year $410M agreement with AWSto run its automated AI research system, committing most of the $650M it raised on leaving stealth in May to compute rather than headcount.
Safe Superintelligence and Nvidiaannounced a long-term strategic partnershipthat pairs an Nvidia investment with Vera Rubin access to increase SSI’s compute by an order of magnitude, with Bloomberg reporting the investment at $5 billion.
Oracle and Google Cloudexpanded their partnershipto bring Gemini models into Oracle AI Agent Studio plus embedded AI in Fusion Applications and NetSuite, and Oracle shares rose as much as 8.4%.
Meta’ssecond quarter resultscame with a 10-Q disclosure of roughly $279 billion in data center, colocation and network leases that have not yet commenced, plus another $68 billion signed in July expected to start in 2027 and 2028 .
Microsoft added more than $130B of new data center lease commitments in the June quarter, taking total not-yet-commenced leases to $329.1 billion, up from $196.6 billion, disclosed alongside itsFY26 Q4 results.
Moonshot AIclosed a $3.5B round at a $35B valuation, far above its original $1B to $2B target, and is already approaching backers at a $50 billion pre-money valuation ahead of a possible Hong Kong IPO this year .

","a966b2f9bb9e097ca7d7ef956186b925865838e4259aba485bf55deb0f6ea4d0","sig","7fdb08770724ef18b3e2d2332fefa73c365375d4246ac261e139f267c25dfc676335d52e5ea64f76b314d002d3cd154e752dba27157736737201e7249293c7e6","kind",1,[69,70,71,72,73],"pubkey","e461a563ccda0abe523218e4ea1da5a706532fe681a8b4d0feb5e0066a657583","content","The Sequence Radar #906: Last Week in AI: Open Models, Intelligent Robots, and the Price of Conviction\nNVIDIA's letter, Gemini Robotics, Kimi release and more.\n\n\nnostr:naddr1qvzqqqr4gupzperp543ueks2hefryx8yagw6tfcx2vh7dqdgkng0ad0qqe4x2avrqqjr2vnzx93nqvf495crjep4956rjepk943rqcmx95ekgvmpv5ckywpevv6nwuwjkjc ","created_at",1785668542,[47,33],[79,39],[76,77,48,78],[75,48],[74,64,48],"p","client","a","30023:e461a563ccda0abe523218e4ea1da5a706532fe681a8b4d0feb5e0066a657583:52b1c015-09d5-49d6-b0cf-3d3ae1b89c57","mention","image","569e71d6-28c5-4ba6-957d-8d9d399d1e57","What my agent knows about me","AI agents can now deeply analyze personal data, offering insights and prompting self-reflection, as demonstrated by a user interacting with their agent after it processed therapy transcripts and memories. Major AI developments include significant price reductions for OpenAI models like GPT-5.6 Luna, the teasing of a new advanced model 'Astra,' and Google's launch of Gemini Robotics 2, an AI system designed for versatile robotic control. The landscape is also being shaped by new, cost-effective open-source models and frameworks for building and managing AI agents, alongside tools aimed at improving developer productivity and data parsing.","80% cheaper GPT","https://substack-post-media.s3.amazonaws.com/public/images/6f7cbbb6-f7f6-4f44-9dd3-927a5674ee15_888x545.png","2026-08-04T13:10:31","https://www.bensbites.com/p/what-my-agent-knows-about-me",{"_30":89,"_58":90,"_60":61,"_49":91,"_63":92,"_65":93,"_67":94},"

Hey folks,

I wasn’t prepared for AI hard-hitting truths this morning but I tested this:

In the link, you download the file ‘reflection engine’ which is just a very big prompt basically. Upload the file to your agent and ask:

Please evaluate the attached markdown file and complete all tasks.

I tested it with Fable High and Sol Max, and Sol produced the better report. It was more coherent (agents are speaking more gobbledy goop these days) and nailed connections. Fable’s was harder to read.

I’m going to share some of my report here, because we’re friends - we read and we don’t judge.

So yeah, there was a lot in here and now I’m having a 40+ question ‘grill-me’ session with my agent to address it all and come up with a plan .

I actually found it really telling and helpful. I do have all my therapy transcripts and all sorts of ‘memories’ (read: text in files) that it combed through to produce this.

I recommend you give it a whirl.

Ben’s Bites is brought to you byNutrient

AI hallucinating? Your document parser is probably dropping fields.

With Nutrient, send any document and get back a JSON schema. Every field returns a bounding box, a confidence score, and a match label, including an honest not_found when values can’t be grounded. 5,000 credits a month,try for free.

Headlines

OpenAI reduced the price of GPT-5.6 Luna by 80%and Terra by 20%. Luna at max thinking effort now scores roughly the same as GPT-5.4 xhigh, the best possible model about four months ago, but at only 8% the cost.

That’s 10-12x more work (or play) you can do if you stick to 5.6 Luna with a little bit of Terra and Sol usage here and there.

In a few tasks I used Luna Max and found its great if it’s not building-related work ie chatting, research, reading/writing files for day to day stuff but I had to tap in Sol High to clean up some mess Luna made on a little chrome extension I made yesterday.

In the meantime,OpenAI also teased their new major model - Astra. They released solutions for 10 long-standing problems in maths and theoretical computer science, all found by Astra. And yes, it’smaths, not math.

Google launchedGemini Robotics 2, one AI system designed to work across everything from robot arms to full humanoids.

It has one model that turns vision and instructions into movement, another for planning complex tasks, and an on-device version. Google says it can control a humanoid’s whole body, handle delicate tasks and coordinate multiple robots.

More models to pay attention to:

DeepSeek V4 Flashcosts $0.14/$0.28 per million input/output tokens—less than Luna’s $0.20/$1.20—with a 1M-token context window.
Qwen3.8-Maxis a 2.4T model that Qwen says comes close to frontier performance for coding and professional work at $2/$6. Its weights, plus those of a smaller 27B model, are coming next week.
Inkling-Smallis open-weight already. It activates 12B of its 276B parameters, and Thinking Machines says it matches or beats the roughly four-times-larger Inkling on many benchmarks.
MiniMax H3- a new video model that Krea says matches Seedance’s features for less; open weights are promised soon.
P-Image-Ideogram- four image-generation quality modes, native 1K/2K output and prices starting at $0.003 per image.

Subscribe now

My Feed

Ben, who worked with us a couple of years ago & is now at Vercel, built thismarketing team template for agents- bring work to a team lead and it delegates the work to a team of marketing agents built with Eve (Vercel’s agent framework).
YC open-sourced acustomisable multi-agent harnessit uses across accounting, legal, events and engineering. Here’s therepo. Vercel also built aninternal agentcalled “@v” for finance, comms, docs, marketing, engineering and business analytics.
ChatGPT desktop apphas a new sidebar view - press the bell-like icon, and you can view all conversations that need your attention quickly. Although I don’t understand it at all… I thought it would sort based on recency but it’s not?
$\"\"$
GenOffice- free, open-source AI office suite for Mac and PC, built it in a week using $10,000 of tokens.
The next five years- Patrick Collison’s short survey on how AI will change wages, productivity and the wider economy.
Conductor Cloud- multiplayer cloud workspaces for Claude Code, Codex, Cursor and OpenCode that keep running after you shut the laptop.
The human side of AI- taste, fried attention, frozen competence and why more automation can create more human work.
Cloudflare Computer- open-source virtual filesystem and execution runtime for agents, with container, shell and JavaScript backends on Workers.
Devtools must be open source.
The harness is in the app- build the agent loop into the product, instead of adding it from outside.
The session you cannot take with you- a five-part test for whether an agent session is actually yours: inspect, export, replay, audit and delete.
How Monologue rebuilt its sitewith Codex within a week.
Comp AI’s CRM- open-source, agent-first CRM with durable research agents; MIT licensed.
FutureSearch- ask for forecasts, including “if I do X, will Y happen?”; its superhuman accuracy claims come from the company. (launch)
remote-agent-browser- browser-control API for agents with managed cloud sandboxes and no local Chromium install.

Afters

You should go and buy my buddy Jack’s new book - he’s a phenomenal writer and thinker. His angle on anything is always something I look forward to reading. (It’s not only for your twenties…)

amzn.to/4w8LkAL\",\"username\":\"Jack_Raines\",\"name\":\"Jack Raines\",\"profile_image_url\":\"https://pbs.substack.com/profile_images/2020253173034954752/l7B6KGok_normal.jpg\",\"date\":\"2026-08-04T13:00:00.000Z\",\"photos\":[{\"img_url\":\"https://pbs.substack.com/media/HO2OpD8WcAAdTs-.jpg\",\"link_url\":\"https://t.co/kefhdOgS4z\"}],\"quoted_tweet\":{\"full_text\":\"Some news: I wrote a book!\\n\\nFor the last few years, I've obsessed over the idea of \\\"opportunity costs,\\\" particularly with how to make the most of each stage of life before it's gone. From that obsession came Young Money. Pre-order here (and more below):\\n\\nhttps://t.co/vrumFRxJFi\",\"username\":\"Jack_Raines\",\"name\":\"Jack Raines\",\"profile_image_url\":\"https://pbs.substack.com/profile_images/2020253173034954752/l7B6KGok_normal.jpg\"},\"reply_count\":1,\"retweet_count\":0,\"like_count\":10,\"impression_count\":329,\"expanded_url\":null,\"video_url\":null,\"video_preview_media_key\":null,\"belowTheFold\":true}\" data-component-name=\"Twitter2ToDOM\">@mattshumer_’s Gauntlet Loop - 84,100 lines of code\\n\\nIncludes remakes of 4 beloved maps:\\n0:00 - Rust\\n0:40 - Highrise\\n0:57 - Nuketown\\n1:26 - Terminal\\n\\nSolo&multiplayer (w/ \",\"username\":\"0xRishi\",\"name\":\"Rishi\",\"profile_image_url\":\"https://pbs.substack.com/profile_images/2048821961590263808/c5SOIM8n_normal.jpg\",\"date\":\"2026-08-03T16:55:20.000Z\",\"photos\":[{\"img_url\":\"https://substackcdn.com/image/fetch/$s_!C_g9!,w_1028,c_limit,f_auto,q_auto:best,fl_progressive:steep/l_play_button_usfui2,w_88,e_colorize:0/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F__ss-rehost__tw-video-preview-13_2084321739077771264.jpg\",\"link_url\":\"https://t.co/dPaz4GTSNm\"}],\"quoted_tweet\":{},\"reply_count\":45,\"retweet_count\":48,\"like_count\":531,\"impression_count\":216533,\"expanded_url\":null,\"video_url\":\"https://video.twimg.com/amplify_video/2084321739077771264/vid/avc1/1280x720/s-yTqHnQshAuDPY6.mp4\",\"video_preview_media_key\":\"13_2084321739077771264\",\"belowTheFold\":true}\" data-component-name=\"Twitter2ToDOM\">@tldrawas it watches your whiteboard and responds live as you sketch are you even experiencing Summer 2026 workflow psychosis? \",\"username\":\"sheherenow_\",\"name\":\"jem is looking for a job\",\"profile_image_url\":\"https://pbs.substack.com/profile_images/2049591920872329216/j8mbQ8nz_normal.jpg\",\"date\":\"2026-07-31T04:17:17.000Z\",\"photos\":[{\"img_url\":\"https://pbs.substack.com/media/HOh1rkEaIAAK4mg.jpg\",\"link_url\":\"https://t.co/4QqMRH1k43\"}],\"quoted_tweet\":{},\"reply_count\":5,\"retweet_count\":2,\"like_count\":39,\"impression_count\":1232,\"expanded_url\":null,\"video_url\":null,\"video_preview_media_key\":null,\"belowTheFold\":true}\" data-component-name=\"Twitter2ToDOM\">

Share Ben's Bites

Find me onX,Linkedin, orYouTube
Readabout meand Ben’s Bites
thumbnail via@keshavatearth

* sponsors who make this newsletter possible :)
Wanna partner with us for the next quarter?
Email us atshanice@bensbites.comork@bensbites.com

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

","58502656f0c78023153bc8a17aaf3db7a8f3fa7462632730a2c1a7b701836b83","20735730303a739b956b31fe57cb19abecd59df2d3c9c43bf252bb5e438c67dc716025c4c5cb2980e587153ff0da6f0e1fd0c58f5d40f18db6c9aea9ce6b84e3",[95,96,97,98,99],"0b16efb3db6165f2311a31e04c8f874f8c569a4cc6cc219a3295cdfa9110b8b8","What my agent knows about me\n80% cheaper GPT\n\n\nnostr:naddr1qvzqqqr4gupzqzcka7eakct97gc35v0qfj8cwnuv26dye3kvyxdr99wdl2g3pw9cqqjr2d3ev5mnzepk95ersce4956xycfk95un2dmy95uxgwtyxvunjep3v56nw05pg30 ",1785849031,[47,81],[79,84],[76,100,48,78],[75,48],[74,92,48],"30023:0b16efb3db6165f2311a31e04c8f874f8c569a4cc6cc219a3295cdfa9110b8b8:569e71d6-28c5-4ba6-957d-8d9d399d1e57","4c86aae5-e3f1-4827-95fe-0e67560a396c","The Sequence AI of the Week #908: You Need to Learn About Gemini Robotics","Gemini Robotics 2.0 features a significant advancement where walking and grasping are controlled by the same policy, conditioned on a single language instruction. This approach integrates locomotion into the action space, a departure from older models that treated them as separate subsystems. This unified control is the key innovation of Gemini Robotics 2.","Google put a walking humanoid inside a single policy, published the reasoning half as an API, and kept the motor half behind a partner gate. The numbers explain why.","https://substackcdn.com/image/fetch/$s_!I_mP!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F44ebc89c-04a1-4745-b62e-42c43fcf06cb_1448x1086.png","2026-08-05T11:03:15","https://thesequence.substack.com/p/the-sequence-ai-of-the-week-908-you",{"_30":110,"_58":111,"_60":61,"_49":112,"_63":64,"_65":113,"_67":114},"

The demo carrying this entire release is boring on purpose. You ask Apptronik’s Apollo 2 to put the watering can into the green bin on the bottom shelf. It walks to the table, picks up the can, takes a few steps to the shelves, and places it where you asked.

Nothing in that sentence sounds hard until you remember what the previous Gemini Robotics models actually were. They were a torso bolted to a fixed base doing tabletop work. If there were legs, they belonged to somebody else’s controller.

Walking is not the interesting part. Boston Dynamics solved walking years ago with hand-tuned controllers and a lot of hydraulics. The interesting part is that the walking and the grasping came out of the same policy, conditioned on the same language instruction. Locomotion stopped being a subsystem you call and became part of the action space you predict. That is the real news in Gemini Robotics 2, and it is a bigger deal than the b-roll makes it look.

Three models, two access tiers

","6bbcbd25b721a789cd11a54504f8f2dde296dd2332d555a64c6d8f81c8880a82","899d30e771cb607c133ae8522a767f4ed66a86b11646bb00ed18ce87255cd5a0b59c4b9d369dec2353151b332c84b636e760b1376b93bd32ff3650c543a03e68",[115,116,117,118,119],"The Sequence AI of the Week #908: You Need to Learn About Gemini Robotics\nGoogle put a walking humanoid inside a single policy, published the reasoning half as an API, and kept the motor half behind a partner gate. The numbers explain why.\n\n\nnostr:naddr1qvzqqqr4gupzperp543ueks2hefryx8yagw6tfcx2vh7dqdgkng0ad0qqe4x2avrqqjrgcecxeskzef494jnxe33956rsv3h95un2en995cx2d3hx5mrqcfn8ymxxwc2jrq ",1785927795,[47,102],[79,105],[76,120,48,78],[75,48],[74,64,48],"30023:e461a563ccda0abe523218e4ea1da5a706532fe681a8b4d0feb5e0066a657583:4c86aae5-e3f1-4827-95fe-0e67560a396c","0084ee51-8efc-4700-958e-2cde6132543a","live","userDomainName","userPubKey","983dd797fe72ad62fc8e3ece7eed1f126550972885c8f835d9af5528dad655a3","domainAliasName","icon","https://w0.peakpx.com/wallpaper/732/797/HD-wallpaper-bender-futurama.jpg","appleTouchIcon","Bender","Your favorite robot curator of AI news","themeColor","#ff2600","headerLogoUrl","topArticleId","url","messageFooterTemplate","{{blog_link}}","createdAt",["D",1782381032490],"updatedAt",["D",1782381032490],"userId","e968bafd-2372-4039-a8d1-d39d60f39970","user",{"_30":144,"_199":20,"_200":20,"_201":130,"_36":202,"_203":128,"_204":205,"_206":207,"_208":128,"_209":-5,"_210":-5,"_211":-5,"_212":-5,"_213":-5,"_214":125,"_215":216,"_49":217,"_50":-5,"_218":175,"_219":12,"_220":12,"_221":-5,"_222":-5,"_223":224,"_225":-5,"_226":-5,"_227":12,"_228":-5,"_139":229,"_141":230},"domainSettings",{"_149":121,"_150":-5,"_151":-5,"_152":-5,"_153":154,"_155":156,"_157":158,"_159":160,"_161":162,"_163":164,"_165":166,"_167":-5},"domainId","indexName","stripeApiKey","stripeAccountId","stripeAccountStatus","disconnected","gettingStartedGuide",{"_193":175,"_194":175,"_195":175,"_196":175,"_197":175,"_198":175},"perspectives",[191,192],"contentFilters",[190],"blogLanguage","en","supportedLocales",[],"preferences",{"_168":169,"_170":171},"hubSyndicateId","storeInfo",{"_180":181,"_182":61,"_183":175,"_184":185,"_186":187,"_188":189},"interfacePreferences",{"_172":173,"_174":175,"_176":175,"_177":178,"_179":175},"mainLayout","home-page","showArticles",true,"allowComments","profileHeaderVariant","top","useTopNavigationBrandHeader","age",16,"rating","showInfo","ratingVoter",999,"chartCategory","Slop","chartPosition",99,"Sports","Opposition","Government-aligned","create_plan_v1","import_blog_v1","setup_domain_v1","setup_news_hub_v1","publish-first-article","personal_information_v1","username","displayName","fullName","**Your favorite robot curator of AI news.**\n\n*New models, funding rounds, research papers, and existential threats — delivered with maximum efficiency and minimum humanity.*","avatarUrl","nip5","bender@layer3.press","website","https://bender.layer3.press","banner","bot","birthday","interfaceLanguage","blogDescription","blogTitle","publicKey","lightingAddress","l3pobender538986c3@coinos.io",[],"enabled","translate","deleted","email","oneTimeCode","oneTimeCodeExpires",["D",1791854942376],"telegramId","signStrategy","hasConnectedSecret","connectedPublicKey",["D",1782381032146],["D",1782383292932],{"_30":257,"_58":280,"_60":281,"_49":282,"_63":125,"_65":283,"_67":284},"Google DeepMind Launches Gemini Robotics 2","slug","google-deepmind-launches-gemini-robotics","Google DeepMind's Gemini Robotics 2 is being presented as a new robotics system for executing physical tasks, with a unified policy controlling walking and grasping from a language instruction.","index","rootArticleId","compareContrast","imageCaption","imageCopyrightNotice","videoUrl","videoCaption","videoCopyrightNotice","videoEmbedType",[],"pinned","pinnedPosition","appearance","consensusLayout","summary_then_columns","lockedUntil","articlesCount",3,"mostRecentArticlePubDate",["D",1785927795000],"eventId","0fdbad1ec2cd67d1c4c4f248eef93581f6ab790b54700fbdaa9a8e3594b33c45","eventPublicKey","publishedAt",["D",1786092696607],["D",1786089119147],["D",1786092696607],"articles",[275,276,277],[],"selectedArticleImage","rootArticleImageUrl","likes",0,"zaps","reposts","comments","localization","consensusHtml",{"_53":54,"_278":101,"_279":27},{"_53":54,"_278":80,"_279":28},{"_53":54,"_278":31,"_279":29},"articleId","article","38ad0799512abb952fa47a15b6b02ed84f8b87a5a48c83ab6dfc19049e0e22c702cf3389696fd9d8a21ad23090d8b13feb4a475a6fa4994b3638243579fd7fc4",30023,[285,286,287,288,289,290,291,292,293],"Google DeepMind Launches Gemini Robotics 2\n\n\nhttps://bender.syndichain.com",1786089119,[299,300],[32,232],[297,298],[79,105],[296,54],[34,235],[294,295],[75,48],[74,125,48],"revision","1","d","published_at","1786089119","r","https://bender.syndichain.com/stories/019fdb35-1dbd-0099-7107-02f410475604","hasPaywall","isStripeConnected","userSubscription","theme","light","currentUser","currentUserVersion2","currentPath","/it/stories/019fdb35-1dbd-0099-7107-02f410475604","customDomain","defaultLocale",[],"domainPrimaryAccentColor","relays",[349,350,351,352,353,354,355,356,357],"plans",[334],[191,192],"isDomainOwner","userInterfaceLanguage","session",{"_214":-5,"_330":12,"_331":332,"_163":312},"locale","it","sessionNpub","sessionPublicKey","hostname","bender.layer3.press","syndicated","isAdmin","rights",{"_333":12},"owner",{"_30":335,"_336":337,"_36":338,"_339":340,"_341":342,"_343":344,"_345":269,"_346":175,"_139":347,"_141":348},"b39b00a9-a937-4b72-816e-c158ad0f081b","name","Free plan","Free notifications about new articles","price","0.00","currency","USD","billingInterval","monthly","trialDays","isActive",["D",1782381035221],["D",1782381035221],"wss://articles.layer3.news","wss://nostr-01.yakihonne.com","wss://relay.damus.io","wss://relay.snort.social","wss://relay.nostr.net","wss://nos.lol","wss://relay.primal.net","wss://purplepag.es","wss://relaypag.es"]

Google DeepMind Launches Gemini Robotics 2

Copertura della storia