|
|
|
So this might be the most heavy metal thing weāve seen all day (pun heavily intended): the worldās first actual mech suit transformer. |
|
Itās a giant, pilotable mech suitā¦ essentially every iconic Japanese anime youāve ever seen from Gundam to Neon Genesis Evangelion is about to become a reality. |
But not ONLY that. It can knock over walls, and terrifyingly, it can turn into a giant quadruped dog. Yāknow, not to chase you down or anything weird like thatā¦ |
This reminds me: back in 2017, there was a big Kickstarter campaign to fund a giant mega battle robot duel that didnāt quite live up to expectationsā¦ but DID happen.Ā |
Hereās what happened in AI today: |
š» Google's Magic Pointer pushed AI beyond prompt boxes. š° Anthropic refused to give China access to its newest model. š° Isomorphic Labs raised $2.1B for drug discovery. šŖ Claude Code added a multi-agent dashboard. š Ask AI to write the assignment first.
|
ā¦and a whole lot more that you can read about here. |
Want to reach 720,000+ AI-hungry readers? Advertise with us!Ā |
P.S: Love robots? Weāre starting a new robotics newsletter! Sign up early here. |
|
š Google Intelligence, Perceptron, and a Whole New World of Interaction |
|
Click here for the FULL BRIEF |
As you probably know, Google has A LOT of AI experiments going on right now. Theyāre historically and to this day one of the most prolific companies building with AIā¦ so much so that the theme of the Musk V. Altman trials happening atm is basically āeveryone V. Googleā (because OpenAI the non-profit was founded to stop them from owning all of AI). |
Well, today Google made its next big bet to own the future of AI, and it looks a little something like this. |
Hereās what happened: |
Google announced Gemini Intelligence for Android, which can automate app tasks, summarize and compare web pages, fill forms, turn messy dictation into polished messages, and build custom widgets. Googlebook is Googleās new premium laptop category, built around Gemini, Android apps, Chrome, phone sync, Magic Pointer, and premium hardware. Wait a minute, whatās Magic Pointer? Wellā¦ Google DeepMind showed off Magic Pointer, a Gemini-powered cursor that understands what you are pointing at and can act on āthisā or āthatā without a full prompt.
|
Of those three, the magic mouse killer is DEFINITELY the most interesting. A normal cursor tells the computer where you clicked. Google wants the cursor to tell the computer what you mean.Ā |
Point at a date in an email, and it can make a meeting.Ā Point at a table, and it can turn the numbers into a chart.Ā Point at a paused travel video, and it can find the restaurant on the map.
|
Thatās the real shift: the interface starts carrying part of the prompt for you. |
And Google is not alone in this.Ā Thinking Machines Lab from ex-OpenAI CTO Mira Murati previewed interaction models, which process audio, video, and text in tiny 200-millisecond chunks so the model can listen, watch, interrupt, and use tools in real time. Meanwhile, Perceptron released Mk1, a model built to understand video as a stream of events, not a pile of screenshots. |
Why this matters:Ā In the Genspark interview weāll release later today, COO Wen Sang told us he made a public commitment to be in front of his computer 50% less by the end of this year. That only works if AI helps us move off screens as the center of our work; and new interaction models make it easier to communicate without having to type or click, meaning eventually we can change the surface area of computers to fit whatever form is most convenient for us without sacrificing functionality.Ā |
Also, everyone is obsessed with being able to control coding agents from your phone now. But the end state of agentic intelligence should not be typing commands into Telegram forever (arenāt yāall sick of texting, or is it just me and Appleās terrible keyboard AI?) It should be the removal of every barrier between you and the technology, so you can interact with it in whatever way best suits your needs. |
This leads me to believe the dream of āambient intelligenceā, or intelligence that is built into the background of our world to support us only when called upon (like āmagicā) could finally become realityā¦ |
|
|
|
|
The Dell AI Factory with NVIDIA delivers a validated, end-to-end AI stack spanning infrastructure, software, and services, designed to help organizations operationalize agentic AI at scale. |
Built for production:
Up to 269% ROI in year one Up to 86% faster AI deployment Flexible deployment across on-prem, edge, and hybrid environments
|
Learn More |
|
|
Want Codex to keep working instead of stopping after one polite little answer? |
Try this trick from @meta_alchemist: ask Codex to write the /goal prompt first. /goal is an experimental Codex CLI feature, meaning itās currently for the terminal version. OpenAI says it helps Codex pursue long-running work toward a clear stopping condition. |
read this session and repo, analyze deeply the exact intent and goals we are looking to achieve here then write me the /goal prompt for this.
make sure to dig into history & docs we have to be 100% clear
if you are not sure about certain parts or wanna ask me a few questions to clarify certain goals further don't hesitate
|
Then paste Codexās answer back in with /goal at the front. |
Outside Codex, use the same idea with ChatGPT or Claude: have it read your context and write the system prompt before it starts your research memo, strategy doc, sales deck, or meeting follow-up. |
Want more in depth tips like this? Check out our AI Skill of the Day Digest for May. |
Have a specific skill you want to learn?Ā Request it here!Ā |
|
|
|
|
|
|
Yesterday, Isomorphic Labs just raised a $2.1B Series B to scale AI-driven drug discovery; if youāre wondering WHY, what they do, or how they plan to do it, check out our recent interview with the Isomorphic Labs team!Ā |
Watch and/or Listen now:Ā YouTube | Spotify | Apple Podcasts |
|
š° Around the Horn |
Anthropic reportedly refused China access to its newest model, while the Pentagon is using the same model family for cybersecurity work. Google and SpaceX are reportedly talking about putting AI data centers into orbit; the pitch is solar power, space cooling, and maybe a way to keep the AI boom from eating every substation in Iowa.Ā Jensen Huang was (maybe??) left off President Trump's China business delegation until he apparently hopped aboard Air Force One in Alaska at the last minute. Mistral and TanStack were hit in a supply-chain campaign that exposed developer credentials across PyPI and npm packages. Meta offered rival AI chatbots one month of free WhatsApp access to address EU antitrust concerns. OdysseyĀ released PROWL, a reinforcement-learning-driven adversarial framework where agents explore game worlds like Minecraft to find world-model failures and feed targeted fixes back into training.
|
|
|
|
Enterprise teams are generating more AI code than ever. Delivery confidence isn't keeping pace. Join engineering leaders at Agentic DevOps World on May 19 for the live data drop on what's actually breaking and how to fix it. |
Ā Register now |
|
š Midweek Wisdom: |
Eric Ries argues mission-driven companies need governance that can outlive the founder, because only 20% of founders remain CEO three years post-IPO and āmission-hopefulā culture usually collapses under investor, board, and acquisition pressure. Tuhin Nair argues AI-era developer teams should ship a fast āSpeedā version for learning, then let seniors harden a āScaleā version for reliability. Beyond Semantic Similarity argues agentic search should inspect source material directly with tools like keyword search and shell commands, because fixed top-k semantic retrieval can throw away useful evidence before the agent gets to reason. Empirical Work in the Age of AI argues AI can 10x empirical research workflows like scraping, replication, and causal analysis, but rigor still depends on humans verifying the data, methods, and weird details models miss. Jie Tang argues the next agent breakthrough is long-horizon work, like models hunting bugs around the clock, as memory, continual learning, and self-judging turn agents into systems that improve themselves.
|
|
|
|
|
| Thatās all for now. | | What'd you think of today's email? | |
|
|
P.S: Before you goā¦ have you subscribed to our YouTube Channel? If not, can you? |
 | Click the image to subscribe! |
|
P.P.S: Love the newsletter, but only want to get it once per week? Donāt unsubscribeāupdate your preferences here. |
