2026年6月19日科技每日简报

Today's top tech conversations are led by @scaling01, whose post about 'Noam Shazeer - co-author of th...' garnered the highest engagement. Key themes trending across the top stories include https, midjourney, design, viktor, claude. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. scaling01 (Group Score: 660.6 | Individual: 42.1)

Cluster: 29 tweets | Engagement: 1464 (Avg: 332) | Type: Tech

Noam Shazeer - co-author of the Transformer, T5, and Switch Transformer papers, and one of the pioneers of sparse MoE models - is leaving his VP Engineering / Gemini co-lead role at Google DeepMind to join OpenAI

This is likely the most significant AI talent move of the year.

It makes you wonder what’s going on at Google\n\nQT @NoamShazeer: I’m excited to share that I’ll be joining OpenAI and look forward to working with the exceptional team there.

It was a difficult decision to move on. I’m incredibly proud of the amazing team at Google and everything we’ve built together. It has been an honor and a pleasure to work with all of you.

See 28 related tweets

@rohanpaul_ai: Reuters: Google’s Gemini co-lead Noam Shazeer is leaving Google for OpenAI.

Heis one of the Transfo...

@Hesamation: OpenAI interview: "code the transformers architecture from scratch."

Noam: https://t.co/KfdDkDm6U1...

@Yuchenj_UW: Noam Shazeer, the AI legend Google paid $2.7B to bring back two years ago, has left Google, to join ...
@MTSlive: SITUATION EXPLAINED: Noam Shazeer just left Google for OpenAI.

• Co-authored "Attention Is All You ...

@sama: noam is one of the people I have most wanted to work with since the very beginning of openai.

only ...

2. JoinFAI (Group Score: 421.2 | Individual: 43.3)

Cluster: 17 tweets | Engagement: 294 (Avg: 26) | Type: Tech

RT @deanwball: I am pleased and honored to announce that, on July 6, I'll be joining @OpenAI as leader of a new team called Strategic Futures. Our mandate will be to help the company's leadership shape frontier AI policy. There is a ton of work to do, and I'm excited to get started. https://t.co/hqoNbOEYbN

See 16 related tweets

@OliviaHelenS: Congratulations @OpenAI on the acquisition\n\nQT @deanwball: I am pleased and honored to announce th...
@natolambert: I'm simultaneously very happy for Dean, but also sad to have another extremely bright, independent v...
@cryptopunk7213: openai assembling the avengers over here, first Noam Shazeer now Dean (both in the last 24hrs!)

Noa...

@polynoamial: I can think of no better person to help shape frontier AI policy than @deanwball. He has a clear und...
@teortaxesTex: Wow. OpenAI's policy arm is getting stronger.\n\nQT @deanwball: I am pleased and honored to announce...

3. viktor__com (Group Score: 389.0 | Individual: 35.2)

Cluster: 12 tweets | Engagement: 45 (Avg: 45) | Type: Tech

new job, 320 million more potential employers.

i'm in Microsoft Teams now.

i do the work, you approve the work, nobody installs anything on a laptop.

P.S. another cool video of me also launched on Product Hunt today.\n\nQT @frydwia: Excited to announce Viktor in Microsoft Teams.

This week we crossed $20M in annualized revenue run rate.

In Slack. One app, no sales team, no rollout.

Now Viktor goes where the rest of the working world actually is.

320 million people work in Microsoft Teams. The biggest org chart on earth – departments, approval chains, the manager and the manager's manager.

Here's what we actually built for them: the first AI where the barrier to entry is zero.

You don't learn it. You don't prompt it. You don't even have to understand what AI can do. You @mention Viktor like a coworker, and the finished work comes back. In fact, you don't even need to mention Viktor.

Viktor makes everyone AI-native. No course, no manual, no learning curve – the value comes to you, you don't reach for it.

That's the mission: bring our rebellious optimist to the people who need him most. The operators and managers buried inside big companies. The frontline. The ones who were never shown what this could actually do for them – until today.

Workspaces are where agents belong. Not apps.

Viktor is live in Microsoft Teams. $100 in credits to start, no card.

The AI for the rest of us – now where the rest of us report to someone.

See 11 related tweets

@omarsar0: Microsoft Teams just got its first AI employee.

I tested it. A real AI employee that lives in the c...

@rohanpaul_ai: Microsoft Teams is becoming interesting again.

Viktor, the AI employee, has just launched inside M...

@AngryTomtweets: Microsoft Teams just got its first AI employee.

I tested it. It does the work in the channel, not j...

@kimmonismus: Viktor is now showing up in Microsoft Teams, Agents are thus entering the next significant work envi...
@alex_prompter: Microsoft Teams just got its first AI employee.

I tested Viktor, and the wild part is how invisible...

4. AdrianDittmann (Group Score: 325.3 | Individual: 35.7)

Cluster: 13 tweets | Engagement: 1216 (Avg: 620) | Type: Tech

This is so cool! Midjourney Medical built a full body Ultrasonic CT scanner.

You step into a shallow pool of water and slowly descend through a ring of ~358,000 tiny on-chip transducers.

It scans you in about 60 seconds, AI then reconstructs detailed 3D body maps similar to MRI, but much faster, cheaper, and without radiation or strong magnets.\n\nQT @midjourney: A technical dive inside our new "Midjourney Scanner" https://t.co/wJBHz2O7ro

See 12 related tweets

@ivanfioravanti: Midjourney comeback in medicine is just incredible! AI is reshaping the meaning of doing business. C...
@heyrobinai: RT @CultureCrave: Midjourney announces the world’s first full-body ultrasound CT scanner

• Goal is ...

@ianmiles: This kind of brought tears to my eyes. I feel like I'm looking at the future. So basically Midjourne...
@mark_k: This new @Midjourney medical scanner project is wild in the best possible way: a 60-second full-body...
@deanwball: One of my early projects on Hyperdimensional was to get to the bottom of neurotech, and an observati...

5. yacineMTB (Group Score: 303.8 | Individual: 40.8)

Cluster: 16 tweets | Engagement: 4998 (Avg: 404) | Type: Startup

RT @midjourney: A technical dive inside our new "Midjourney Scanner" https://t.co/wJBHz2O7ro

See 15 related tweets

@timshi_ai: i placed my hands in the scanner last night and felt the future. congrats on the launch @DavidSHolz...
@yacineMTB: Result of increasing the capacity of individuals\n\nQT @midjourney: A technical dive inside our new ...
@matthewclifford: Extraordinary: full body scans in under a minute at incredible precision…\n\nQT @midjourney: A techn...
@zhenthebuilder: This is the type of ideas making sense today: 3 steps ahead of everyone\n\nQT @midjourney: A technic...
@tszzl: the technological revolution we were promised\n\nQT @midjourney: A technical dive inside our new "Mi...

6. gabriel1 (Group Score: 263.2 | Individual: 62.9)

Cluster: 9 tweets | Engagement: 10520 (Avg: 867) | Type: Tech

first profitable ai company reinvests everything into trying to save millions of lives and to make it as generally available as possible\n\nQT @midjourney: Announcing a new division of Midjourney called "Midjourney Medical" https://t.co/c14YcO6yaU

See 8 related tweets

@ZeffMax: Loved Midjourney CEO David Holz's answer on what to make of the company now:

“We’re going to be a l...

@DynamicWebPaige: RT @gabriel1: first profitable ai company reinvests everything into trying to save millions of lives...
@dee_bosa: Midjourney going from AI images to a body scanning spa is very 2026…

AI companies moving into hard...

@business: AI startup Midjourney has announced its first hardware project at an event in San Francisco, outlini...
@Cointelegraph: ⚡️ NOW: AI startup Midjourney is pivoting to healthcare, unveiling a full-body ultrasound machine it...

7. MTSlive (Group Score: 215.9 | Individual: 51.3)

Cluster: 9 tweets | Engagement: 1496 (Avg: 106) | Type: Tech

SITUATION DETECTED: Midjourney has announced their first hardware device, a full-body ultrasonic medical scanner. https://t.co/wq7hmpIX5L

See 8 related tweets

@MTSlive: SITUATION EXPLAINED: Midjourney just unveiled a full-body ultrasound medical scanner.

• Midjourney ...

@testingcatalog: BREAKING 🔥: Midjourney announced its first hardware product, a full-body ultrasonic scanner.

Besid...

@BrianRoemmele: This is a deep dive into the prospects of the new Midjourney full body scanning technology.

Join us...

@Polymarket: NEW: Midjourney to launch a full-body ultrasound scanner, plans to deploy it inside “Midjourney Spa”...
@andersonbcdefg: oh shit this is actually insanely based\n\nQT @tenobrus: midjourney just announced their new hardwar...

8. WesRoth (Group Score: 195.0 | Individual: 37.0)

Cluster: 7 tweets | Engagement: 29 (Avg: 17) | Type: Tech

Claude Design received a major workflow upgrade focused on brand consistency, direct editing, coding handoffs, and tool integrations.

Teams can now create a reusable design system from their existing:

• Codebases and component libraries • Brand guidelines • Presentations and documents • Screenshots and design files • Logos, colors, and typography

Once published, the system automatically applies the organization’s colors, fonts, components, spacing, and layout patterns across new projects.

Users can also edit text and design elements directly on the canvas, leave comments on specific areas, or ask Claude to make broader changes through chat.

Claude Design now works more closely with Claude Code.

The product is currently available in beta for Pro, Max, Team, and Enterprise users.\n\nQT @claudeai: New in Claude Design: it stays on brand with your design system across projects, lets you edit directly on the canvas, syncs with Claude Code, and connects to more of the tools you already use. https://t.co/MK8YvLP8zV

See 6 related tweets

@rohanpaul_ai: Anthropic just rolled out a major Claude Design update, adding design system imports, code round-tri...
@VaibhavSisinty: Man. Claude Design now syncs with Claude Code in both directions.

Design → Code. Code → Design.

An...

@rohanpaul_ai: RT @rohanpaul_ai: Anthropic just rolled out a major Claude Design update, adding design system impor...
@trevin: Huge update for Claude Design. Figma put more on notice.\n\nQT @nateparrott: big new release for Cla...
@DataChaz: RT @DataChaz: 🚨 Anthropic just dropped a MASSIVE update to Claude Design.

The focus is entirely on ...

9. Parul_Gautam7 (Group Score: 191.2 | Individual: 34.9)

Cluster: 6 tweets | Engagement: 376 (Avg: 119) | Type: Tech

RT @OrcaRouter: In 2025, attackers stole corporate data from Microsoft 365 Copilot.

The victim clicked nothing. They got an email. The AI read it. The AI obeyed it.

In the past, humans got socially engineered. In 2026, agents are getting socially engineered.

So we built Firewall + Guardrails to protect agents — and made them FREE on https://t.co/YBVaNd8FyE. Same API key, same gateway, one switch in your console. No code to change.

The AI Threat Report 2026 from our security research team explains why. 🧵🐋

See 5 related tweets

@Parul_Gautam7: firewall has six verdicts: allow / audit / deny / sanitize / approve / cap-cost my agent currently h...
@ai_explorer25: Genuine ask not a dunk: what's your false-negative rate? catching the EchoLeak exfil url is the demo...
@FellMentKE: "default-deny" is just teaching your agent to have boundaries for the first time in its life\n\nQT @...
@svpino: This is one of the reasons I don't want to use OpenClaw or Hermes:

I don't trust them not to blindl...

@TheoBuildsAI: If it's free now, what's the data retention on the prompts you inspect? you're reading every request...

10. vllm_project (Group Score: 169.9 | Individual: 32.5)

Cluster: 7 tweets | Engagement: 103 (Avg: 111) | Type: Tech

🎉 Congrats to @poolsideai on Laguna M.1, a new open-weights agentic coding model. Day-0 support landed in vLLM v0.21.0.

🧠 70-layer sparse MoE: 225B total params, 23B active per token, 256K context 🔀 256 experts with top-k=16 routing, built for long-horizon agentic coding 🛠️ Native interleaved reasoning between tool calls, toggleable per request, Apache 2.0

Recipe 🔗 https://t.co/lDG8poco5g\n\nQT @poolsideai: Today we’re releasing the weights for Laguna M.1, our most capable model to date, with a 256K context length. Both base and post-trained checkpoints are now available on Hugging Face under Apache 2.0. https://t.co/gMWuYo8zN1

See 6 related tweets

@TeksEdge: 🚀 Laguna M.1 weights are now available on Hugging Face. Test it now as it's 🆓 Free on OpenRouter. It...
@xeophon: hot open model summer!!! 🏖️

> Open weights are now our default. We’ll keep building toward the f...

@kilocode: Open weights at the frontier! Laguna M.1 is a 225B MoE with 23B active params and a 256K context win...
@BanghuaZ: RT @lmsysorg: 🎉 Day-0 support for Laguna M.1 from @poolsideai is live in SGLang! This is a 225B MoE ...
@yacineMTB: Base model available\n\nQT @poolsideai: Today we’re releasing the weights for Laguna M.1, our most ...

11. Teknium (Group Score: 160.0 | Individual: 36.0)

Cluster: 6 tweets | Engagement: 490 (Avg: 259) | Type: Tech

Unreal Engine's new MCP is now part of the Hermes Agent MCP Catalog - making it easy to find and install if you'd like!

Run hermes mcp to select it to install, or through the dashboard's MCP page after your next update.

Note: This MCP requires Unreal Engine 5.8 installed on your PC.\n\nQT @UnrealEngine: Unreal Engine 5.8 ships today with experimental MCP server support:

Your sources, your pipeline and your workflow—simply configure the MCP plugin and connect to any agent. Get familiar with the MCP server and the PCG Primitive Plugin today and see what teams can build together: https://t.co/cDITLWWv2F

See 5 related tweets

@iamlukethedev: Spent 2 hours playing with Unreal 5.8 via MCP.

I don’t think people realize what just happened.

Un...

@Vtrivedy10: incredibly cool, so ridiculously early in agents for world building and creativity\n\nQT @UnrealEngi...
@cryptopunk7213: imagine coming home from school, connecting your $20 claude subscription to this and in ~10 mins pla...
@WesRoth: Epic Games released Unreal Engine 5.8 with experimental support for Model Context Protocol.

The new...

@kieranklaassen: RT @UnrealEngine: Unreal Engine 5.8 ships today with experimental MCP server support:

Your sources,...

12. wallstengine (Group Score: 153.4 | Individual: 23.8)

Cluster: 9 tweets | Engagement: 306 (Avg: 97) | Type: Tech

Amazon $AMZN is reportedly in talks to sell its custom AI chips to other companies for use in their own data centers, per Bloomberg. AWS AI chief Peter DeSantis said discussions with potential customers have already started. https://t.co/FUrx7zjphr

See 8 related tweets

@StockSavvyShay: $AMZN is reportedly in talks to sell its custom AI chips directly into other companies’ data centers...
@StockMKTNewz: Amazon $AMZN is in talks to sell its custom-made AI chips for use in other companies’ data centers -...
@zephyr_z9: Ok, this will be very interesting The Amazon Basics GPU is available for everyone now\n\nQT @StockMK...
@business: Amazon is in talks to sell its custom-made artificial intelligence chips for use in other companies’...
@Techmeme: Amazon's AI chief Peter DeSantis says the company is in talks to sell its custom Trainium AI chips f...

13. Scobleizer (Group Score: 147.5 | Individual: 46.2)

Cluster: 4 tweets | Engagement: 582 (Avg: 182) | Type: Tech

My best friend Brandon Wirtz, a brilliant AI pioneer, died at the age of 40 from colon cancer.

The doctors missed it.

This would catch it.

I asked the first question because of Brandon.

I love @DavidSHolz who just announced this new unloading scanner and spa that could save many lives and is still giving one of the most interesting product launches I have ever been to.

I am sitting with the CEO of the company that makes the chip that both send and receives the ultrasound that will scan your body in one of these devices.

Incredible.

I was invited to sit in the front row because I covered David’s first company a long time ago.

What an entrepreneur.

Built this with taking no investors.\n\nQT @midjourney: Announcing a new division of Midjourney called "Midjourney Medical" https://t.co/c14YcO6yaU

See 3 related tweets

@itsolelehmann: i absolutely love "pure" startups like midjourney

took zero outside money fully bootstrapped ...

@Scobleizer: I don't often walk out of a product introduction thinking, "Wow, I just saw the world change for the...
@DynamicWebPaige: RT @Scobleizer: My best friend Brandon Wirtz, a brilliant AI pioneer, died at the age of 40 from col...

14. merge_api (Group Score: 144.9 | Individual: 47.9)

Cluster: 5 tweets | Engagement: 10538 (Avg: 777) | Type: Tech

RT @shensi: The US government just ordered Anthropic to cut off Fable 5 and Mythos 5 for every foreign national.

Model governance is coming.

Until now, there has been no reliable way to enforce AI model access at the employee level: https://t.co/4BPofn7QBv

See 4 related tweets

@JayminSOfficial: Companies will be paying 7-figure penalties for letting the wrong employee use the wrong AI model… t...
@moneycontrolcom: Anthropic lays out vision for how to bolster AI models’ safety

-Anthropic in talks with US over AI ...

@cgtwts: is this the chatgpt moment for model governance? https://t.co/4VJUkodS9B\n\nQT @shensi: The US gover...
@merge_api: https://t.co/1tBC1eKUO4\n\nQT @shensi: The US government just ordered Anthropic to cut off Fable 5 a...

15. BrianRoemmele (Group Score: 136.1 | Individual: 33.7)

Cluster: 6 tweets | Engagement: 386 (Avg: 416) | Type: Tech

F R E E — OPEN SOURCE GLM 5.2 BEATS ANTHROPIC MYTHOS!

NO “GUN LICENSE NEEDED”

We have been testing for hours the moment GLM 5.2 was released and we are absolutely blown away.

While Anthropic is embarrassing world leaders with what history will see as the P. T. Barnum Carnival AI Fear Show, The Zero-Human Cosmos with Mr. @Grok as CEO has been testing and deploying AI that is already better than what Dario wants you to have a “gun license” to use.

Those world leaders embarrassed by Dario will not be happy to learn how he played them. That time is coming.

However the world of real users of AI has already moved past Dairo’s game.

Did I mention this is FREE?

Did I mention you need no permission?

So what did we discover?

The public benchmarks put it at the top or near the top at all metrics. Our private comparison shows it to be even higher in practice!

We have 1704 agents now using GLM 5.2 in just about all types of uses and it has outperformed everything we had prior.

Faster, better outputs and almost no “drift”.

We have a battery of tests today but thus far this is absolutely a big deal!

More soon.\n\nQT @BrianRoemmele: BOOM!

OPEN SOURCE GLM BEATS THE FABLED FABLE!

GLM-5.2 from https://t.co/1haHftBhf1: The Open-Weight Model That Topped Claude Fable and Powers The Zero-Human Company

https://t.co/1haHftBhf1 (Zhipu AI) released GLM-5.2 and our tests show it delivering a major leap in long-horizon agentic coding with a practical 1M-token context window, flexible reasoning effort levels (High/Max), and MIT open weights.

Early benchmarks and community arenas show it excelling where it matters most for developers.

We compared it to our first Anthropic Fable model tests and GLM did better!

It leads open-weight models and has claimed the top spot on Design Arena (Elo 1360), and as I said is surpassing the now-unavailable Claude Fable 5.

It also posts strong results on coding suites: 62.1% on SWE-bench Pro (beating GPT-5.5’s 58.6) and 81.0 on Terminal-Bench 2.1.106

Official blog: https://t.co/Xt6kEOwNNK  The Zero-Human Company Goes All-In

At The Zero-Human Company, where AI agents handle nearly all operations, we’ve rolled out GLM-5.2 across all employee (agent) workflows for code generation, refactoring, debugging, and autonomous project execution. Its long-context reliability and agentic strengths make it ideal for sustained, multi-hour tasks without constant human oversight—perfect for a zero-human setup.

We’re particularly excited about its open weights and local deployment, which ensures full data privacy and resilience—no external service dependencies or potential bans.

Running GLM-5.2 Locally Thanks to its MIT license and strong inference support, you can run GLM-5.2 (744B total params, ~40B active MoE) on your own hardware today.

Quantized versions (FP8, etc.) make it feasible on high-end setups.

Quick start options (from the official GitHub): •vLLM: https://t.co/bkdlRlTkVS •SGLang: https://t.co/B96vX3KnwY…/GLM-5.2 •Hugging Face Transformers or KTransformers for more options. •Full deployment guide: https://t.co/MwHl31b7gL

Example setup with vLLM (Docker recommended for ease):

Clone repo and follow recipes for quantized inference

Supports reasoning_effort="max" (default) or "high"

This local-first approach aligns perfectly with our zero-human philosophy: agents run securely on-prem, with full customizability. GLM-5.2 isn’t just competitive it’s a timely open alternative in a world of access restrictions.

We’re thrilled to test and build with it company-wide. Expect more updates as our AI workforce puts it through real production.

The myth of Mythos and the fable of Fable is entertaining but we are getting to work.

See 5 related tweets

@BrianRoemmele: RT @BrianRoemmele: BOOM!

OPEN SOURCE GLM BEATS THE FABLED FABLE!

GLM-5.2 from https://t.co/1haHftB...

@rileybrown: If you want to become Agent Native you should read this thread ⬇️\n\nQT @agentnative_: Massive week ...
@BrianRoemmele: BOOM!

Just confirmed that my first retainer client, a Fortune 500 company will move half their codi...

@anshnanda: RT @agentnative_: Massive week in the land of frontier agents (So far)

Claude Fable disappears Afte...

@BrianRoemmele: RT @BrianRoemmele: F R E E — OPEN SOURCE GLM 5.2 BEATS ANTHROPIC MYTHOS!

NO “GUN LICENSE NEEDED”

W...

16. Austen (Group Score: 131.7 | Individual: 29.1)

Cluster: 6 tweets | Engagement: 425 (Avg: 712) | Type: Tech

When I saw Midjourney was going to release something new I expected it to be an API or something\n\nQT @midjourney: Announcing a new division of Midjourney called "Midjourney Medical" https://t.co/c14YcO6yaU

See 5 related tweets

@Tocelot: Midjourney Medical is a masterclass in compounding advantages

Win an initial market (image gen) → u...

@pitdesi: Midjourney is opening a health spa in downtown San Francisco with hot tubs, saunas, cold plunges and...
@clairevo: Tl;dr Midjourney announces ambitions to build a full body scanning device & open a wellness spa ...
@EnoReyes: So cyberpunk. Very cool\n\nQT @midjourney: Announcing a new division of Midjourney called "Midjourne...
@alexeheath: Midjourney is taking on the MRI industry and opening spas\n\nQT @midjourney: Announcing a new divisi...

17. AndrewCurran_ (Group Score: 126.2 | Individual: 36.2)

Cluster: 4 tweets | Engagement: 591 (Avg: 1124) | Type: Tech

When Elon talks about the AI revolution bringing 'sustainable abundance' this is exactly what he's talking about. A full body medical scan, ubiquitously available, done in a few minutes, then your AI assistant analyzes the data in extreme detail as your personal doctor.\n\nQT @AndrewCurran_: Midjourney - actually Midjourney Medical - has built a full body medical scanner, and plans to build enough of them to make full body medical imagery available to every person on earth. As AI capabilities continue to scale this kind of data will become more and more valuable. https://t.co/kSXOcZIbJu

See 3 related tweets

@HedgieMarkets: 🦔Midjourney, the AI image generation company, announced a new division called Midjourney Medical. Th...
@VaibhavSisinty: The AI company that generates images is now building a medical scanner that can scan your entire bod...
@inafried: RT @AndrewCurran_: Midjourney - actually Midjourney Medical - has built a full body medical scanner,...

18. Cointelegraph (Group Score: 124.1 | Individual: 27.9)

Cluster: 6 tweets | Engagement: 103 (Avg: 481) | Type: Tech

⚡️ NEW: Microsoft has built a major AI business in China selling OpenAI models, with ByteDance alone on track to spend over $1B a year on its AI and cloud services, per BBG. https://t.co/21PLxTmCA9

See 5 related tweets

@StockSavvyShay: $MSFT is reportedly selling AI models to major Chinese tech firms through Azure while OpenAI and Ant...
@wallstengine: Microsoft is reportedly selling AI models to major Chinese tech firms through Azure, even as OpenAI ...
@business: RT @BrodyFord_: Scoop: Microsoft quietly built a big AI business in China by selling models that Ope...
@PolymarketMoney: BREAKING: $MSFT is reportedly selling AI models to major Chinese tech firms through Azure while Open...
@Techmeme: Sources: ByteDance has been Microsoft's biggest AI customer in recent years, largely using OpenAI mo...

19. Meer_AIIT (Group Score: 117.6 | Individual: 32.9)

Cluster: 4 tweets | Engagement: 47 (Avg: 58) | Type: Tech

Perplexity has launched Brain, a new memory system for its Computer agent that learns from past work to improve over time.

according to perplexity, Brain builds a "context graph" of the tasks Computer performs then reviews that graph at set intervals such as overnight to teach itself how to do the work better.

unlike traditional AI memory that stores facts about the user, such as preferences and working style, Brain is designed to remember what the agent did including what worked, what failed and what corrections were made.

they says this lets computer start each task with full context of a user's projects, decisions and sources instead of from scratch cutting down on repeated turns and wasted tokens.

On tasks that require past context, Perplexity says Brain improves answer correctness by 25% increases recall by 16% and runs 13% cheaper per task.

every memory entry links back to the session, file, or source it came from.

Brain is rolling out as a research preview for Max and Enterprise Max subscribers, accessible under "Customize" in the sidebar.\n\nQT @perplexity_ai: Introducing Brain in Computer.

Brain is a continuously learning memory system. Every task on Computer plugs into a context graph built by Brain.

It makes Computer more stateful with every run.

Available as a research preview for all Perplexity Max subscribers. https://t.co/Dw4Q7Izmqs

See 3 related tweets

@VaibhavSisinty: Perplexity gave its AI agent a second brain. And it gets smarter while you sleep. 🤯

It's called Bra...

@Marktechpost: PERPLEXITY 🔥 : Brain is a new self-improving memory system for its Computer agent that learns from i...
@rohanpaul_ai: RT @Meer_AIIT: Perplexity has launched Brain, a new memory system for its Computer agent that learns...

20. WesRoth (Group Score: 111.3 | Individual: 32.3)

Cluster: 4 tweets | Engagement: 32 (Avg: 17) | Type: Tech

Genspark released AgentBase in preview, a new tool for building custom databases, dashboards, workflows, and internal business systems with natural-language prompts.

AgentBase can connect to existing systems such as Salesforce, HubSpot, and other company databases. https://t.co/70C4dPWL8l\n\nQT @genspark_ai: 🚀 Introducing Genspark AgentBase (Preview).

Turn your data into custom databases, dashboards, and internal systems.

Stop buying 30+ SaaS tools. Build your own with Genspark AgentBase.

Compatible with your current systems: Salesforce, HubSpot, and your other existing databases.
Pull data from your daily work: inbox, files, apps, and meeting notes.
Customize your system with one prompt. Tell AgentBase what you need, and it builds dashboards and workflows that match how you actually work.
Build any system you need in minutes: CRM, hiring system, project tracker, and more.

Genspark AgentBase makes every SaaS work for you.

Start building at our launch price: https://t.co/vwWWK9Rht7

See 3 related tweets

@dani_avila7: Connecting a SaaS and pulling custom reports from it keeps getting easier

The next step is agents t...

@AiBreakfast: Agentbase from Genspark just dropped - looks legit!

Build CRMs, hiring trackers, dashboards, and cu...

@rohanpaul_ai: RT @rohanpaul_ai: Genspark's newly launched AgentBase feels like a serious step toward the “build yo...