• AI Valley
  • Posts
  • AI is now “Super Intelligence”

AI is now “Super Intelligence”

PLUS: Google unveils Gemini 4 Argon

Together with

howdy, it’s Barsee again.

happy thursday, AI family, and welcome back to AI Valley.

here are the biggest things worth knowing today:

  • Trump wants the US government to stop calling it AI

  • OpenAI just had a pretty big DevDay

  • Google unveils Gemini 4 Argon

  • Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

SENTIENCE

Courtesy: Sentience

Generic AI gives everyone the same intelligence. Your Sentience is a high-fidelity digital twin shaped by your memories, relationships, voice, and judgment.

Your Sentience connects the context across your email, calendar, messages, meetings, and documents. It uses that context to prepare you for decisions, write responses you would actually send, and act with your approval. It learns from every correction and continually improves.

Your AI should become more like you, not make you more like everyone else.

Create your Sentience with code VALLEY and get one month of Sentience Max free.

*This is sponsored

THROUGH THE VALLEY

1/ Trump wants the US government to stop calling it AI - Trump signed an executive order instructing the US government to refer to AI as “Super Intelligence” (SI) in official communications, websites and policy documents. It doesn’t affect any regulations or contracts that mention AI, but Trump wants the name change “to ensure the American people know this is not your everyday technology.” His top science advisor has 60 days to come up with a proposed federal definition of SI.

Separately, Musk, Zuck, Dario, Jensen, Sundar, Brockman and a bunch of other tech execs signed Trump’s “morally binding” AI safety pact. It’s voluntary for now, with companies agreeing to things like safety testing, external audits and board-level oversight. Trump says actual regulation could come later.

Trump also discussed forming a 10-person committee to oversee the effort and said he’d appoint an AI czar in the coming days.

Though it’s gonna take me a minute to start calling it SI.

2/ OpenAI just had a pretty big DevDay, here’s what’s new:

Dots: your own always-on AI coworkers that run on GPT-6 Astra. They each have their own cloud computer & browser, context, tool access, and can use 4,000+ apps via plugins. You can offload work to a dot, leave, and come back when it’s done or when it needs you. Dots will follow you around and you can use them in ChatGPT, via voice, in Slack/Teams, and soon via text.

ChatGPT Space: a place where you, your team members, and your agents can all work together, access the same files, and share context.

GPT-6.1 Sol: surprisingly close to Astra on coding, computer use, and “professional” tasks, and it’s still $2/$10 per million tokens. 1.05m context window. They’ve also got a new Ultrafast mode, which gets like 300 tokens/sec in Codex.

Codex: can now spin up cloud coding environments so you can code even when your computer is asleep, voice in CLI, parallel agents, code review, and a new security agent that can scan your GitHub repo and prep fixes.

It looks like OpenAI is no longer just building ChatGPT and models, they’re now building agents, computers, workspaces, dev stack, marketplace, etc. etc. It’s a lot.

3/ Google unveils Gemini 4 Argon - their new flagship model built for long-running coding, knowledge work and cybersecurity tasks. The ridiculous bit: it can output up to 1 million tokens in a single run, up from 64K before. It also gets 77.9% on DeepSWE and Google says it leads across a bunch of professional-work benchmarks.

It seems to fall behind a bit on some longer technical and scientific tasks (Astra is better at those) and ML eng/terminal tasks (Opus 5.5 is better at those). Some Googlers have apparently expressed concern that Argon benchmarks better at coding than it actually feels in practice, but Google disagrees.

Argon is rolling out to a small group of trusted cyber defenders first, with paid API customers and AI Ultra getting access later.

Also kinda wild: Google says Argon agents already found optimizations that freed 300+ TiB of memory across its own data centers.

TRENDING TOOLS

  • dots - OpenAI’s always-on agents built to handle everything

  • Grok Bot - it can now hand off coding tasks to Cursor, manage your PRs with GitHub and Origin plugins, and share video demos of what they build

  • Ideogram 4.5 - the most precise image editing model

  • d1 - a decision model like Jev

  • Runway Ads - AI video firm now can connect to a brand’s accounts on Meta, Google, and TikTok

  • Call4me - an ai agent that handles phone calls for you

  • Muse for Small Business - Meta has launched a dedicated agent for small business workplaces

  • Grokpedia - an open, comprehensive collection of all knowledge

  • SideShif - the first AI creator marketer

  • Ava Studio - drop in your website, and it researches, scripts, and produces on-brand videos you can edit or ship

WHAT I'M CONSUMING

THE VALLEY GEMS

What’s trending on social today:

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.