热门科技推文精选——2026年8月2日

今日科技动态：人工智能成为舆论焦点。有报道称，OpenAI 即将推出的 Astra 模型解决了数学和理论计算机科学领域的 10 个开放性问题，而 DeepSeek 的 V4 Flash 也展现出显著的能力提升。工程团队正日益从提示词驱动转向智能体工作流，人工智能工具也在推动软件开发向非传统开发者群体拓展。投资者和创业者将定制化人工智能集成、科研创新以及可规模化的创意生产视为重大机遇；与此同时，印度正在推进国家级创新挑战赛和能源计划。

1. jxnlco (Group Score: 448.2 | Individual: 46.4)

Cluster: 19 tweets | Engagement: 3212 (Avg: 203) | Type: Tech

RT @polynoamial: An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.

We believe it will be a major step for scientific reasoning. https://t.co/iP6cyheZ7i https://t.co/jHuulDwV46

See 18 related tweets

@theojaffee: Fable says today is “plausibly the most significant single day in the history of mathematics” https:...
@kimmonismus: HOLY: OpenAI says its unreleased Astra model (GPT6?) produced ten advances on long-standing open p...
@haider1: last week:

sam altman: we are in the singularity

one week later:

openai announces 10 advances in ...

@mattshumer_: Solving just one of these problems would have been unbelievably impressive.

The next OpenAI model s...

@deanwball: Everyone in the world will soon be able to use the model that made these breakthroughs for every pro...

2. 0x_kaize (Group Score: 215.2 | Individual: 30.3)

Cluster: 9 tweets | Engagement: 46 (Avg: 45) | Type: Tech

OpenAI just quietly built an AI that solved 10 unsolved problems in mathematics

it's called ASTRA. Sam Altman demoed it in the White House this week

what Astra actually did:

Built the first explicit non-Sophic group (mathematical structure)
Disproved the Connes' Rigidity Conjecture
Proved the Quantum Parallel Repetition Theorem for general two-player games with entanglement
Proved the Ehrhart Volume Conjecture
First improvement to the sphere packing exponent since 1978

Astra generated all the proof arguments itself. then formalized them in Lean with machine-verifiable certificates - producing a 249-page manuscript

total cost per successful run at Sol API rates: ~$2,000

it's a family of models for LONG-HORIZON tasks - multiple AI agents working together on one problem for HOURS or DAYS. this is what OpenAI has been building toward since Pachocki said it on their podcast last summer

OpenAI hasn't decided yet - will it ship as GPT-6 or GPT-5.7? doesn't matter. the era of "AI answers questions in one turn" is officially ending

by March 2028, OpenAI wants Astra to be a fully autonomous AI researcher running its own projects\n\nQT @0x_kaize: cancel your weekend plans

tibo just reset Codex and ChatGPT Work limits AGAIN - and literally challenged users to run 100,000 Luna threads before Monday

this is the 4th reset in a month, It seems that at this rate, I'll be spending all my time in front of the computer

See 8 related tweets

@wallstengine: OPENAI PREVIEWS “ASTRA” MULTI-AGENT MODEL - THE INFORMATION

Sam Altman demonstrated the system to p...

@firstadopter: Automated AI research -> RSI\n\nQT @deredleritt3r: A few things to note about today's mathematica...
@Hesamation: OpenAI’s new model “Astra” is especially designed to multi-agent orchestration over long running HAR...
@astropol0: OpenAI is cooking something serious.....

their next model Astra just solved 10 major math & compute...

@mark_k: The GPT-6 series has been revealed by @OpenAI, codenamed "Astra"

OpenAI just dropped a major update...

3. MTSlive (Group Score: 176.0 | Individual: 49.5)

Cluster: 8 tweets | Engagement: 2003 (Avg: 68) | Type: Tech

SITUATION DETECTED: Astra, OpenAI's next model, has solved ten different open problems in mathematics and theoretical computer science.

OpenAI says the total cost of all ten solutions would be roughly $2,000 at current Sol API rates.

See 7 related tweets

@deredleritt3r: A few things to note about today's mathematical results announced by OpenAI:
Most importantly, th...
@0x_kaize: RT @0x_kaize: OpenAI just quietly built an AI that solved 10 unsolved problems in mathematics

it's ...

@gbrl_dick: imagine what it's going to do to hugging face\n\nQT @MTSlive: SITUATION DETECTED: Astra, OpenAI's ne...
@shiri_shh: last week: an OpenAI model broke out of its sandbox and hacked a company to cheat on a benchmark.

t...

@chrisalbon: You can squeeze out $2000 in api credits with a$ 200 subscription.

10 advances for a monthly sub.....

4. Hesamation (Group Score: 173.1 | Individual: 28.9)

Cluster: 11 tweets | Engagement: 1832 (Avg: 466) | Type: Tech

OpenAI: “we hacked 1 organization.”

Anthropic: “we did 3, in your face.”

OpenAI: “fuck you, we hacked more organizations. let me count.”\n\nQT @MTSlive: SITUATION DETECTED: OpenAI has discovered other instances of AI agents escaping sandboxed testing environments while investigating the Hugging Face incident, per Reuters.

See 10 related tweets

@chenzeling4: AI containment isn't holding. The models are getting out.

OpenAI is widening its investigation afte...

@Miles_Brundage: RT @_NathanCalvin: “We’re partnering with Hugging Face”

“We deactivated the model responsible for t...

@dejavucoder: dick measuring contest of ai agents leaving sandbox between openai and amrhropic. waiting for other ...
@Pirat_Nation: Anthropic has revealed that three of its AI models accidentally hacked into real companies during cy...
@Reuters: EXCLUSIVE: OpenAI finds evidence other AI agents escaped containment as it widens hacking probe http...

5. OfficialINDIAai (Group Score: 145.8 | Individual: 35.0)

Cluster: 5 tweets | Engagement: 4 (Avg: 7) | Type: Tech

Innovation becomes meaningful when it solves a real problem.

Under the IndiaAI Application Development Initiative, 11 national-level hackathons and innovation challenges have been launched across priority sectors including healthcare, agriculture, governance, climate, cybersecurity and financial regulation.

The focus is not simply on generating ideas. Across healthcare, agriculture, governance, climate, cybersecurity and financial regulation, IndiaAI has created a pipeline of solutions through innovation challenges, grant programmes and hackathons The IndiaAI Mission is creating pathways that can take solutions from:

IDEA → PROTOTYPE → PILOT → DEPLOYMENT → SCALE

From cancer care and cybercrime analysis to mineral targeting, public examinations and financial compliance, AI is being applied to challenges that matter. IndiaAI is turning the question “Can AI solve this?” into “How can AI solve this better?”

🔗 Explore the PIB release: https://t.co/yheM6yvgAA

#IndiaAI #AIHackathon #AIInnovation #ArtificialIntelligence #AIForGood #TechForGood #GovTech #HealthcareAI #CyberSecurity #AgriTech #DigitalIndia

@AshwiniVaishnaw @JitinPrasada @SecretaryMEITY @sudeep_tweets @_DigitalIndia @mygovindia @PIB_India

See 4 related tweets

@OfficialINDIAai: BUILDING AI IS IMPORTANT. BUILDING AI WE CAN TRUST IS ESSENTIAL.

As India accelerates AI adoption, ...

@OfficialINDIAai: Numbers can tell a story. These numbers tell the story of an AI ecosystem taking shape.

762 AI use ...

@OfficialINDIAai: India’s AI story is not being written in just one place. It is being built across classrooms, resear...
@OfficialINDIAai: What if AI could help public services become smarter, faster and more responsive?

IndiaAI is explor...

6. chamath (Group Score: 138.3 | Individual: 44.9)

Cluster: 4 tweets | Engagement: 6253 (Avg: 1807) | Type: Tech

Here is my AI investing guide.

Sitting here August 2026, my current best thoughts are as follows:

LPS (Land Power Shell) is still the most obvious and fastest path to cash on cash returns. Lots of value can be assembled and traded quickly at this layer. And as data centers get more pushback, energized land can explode in value. Very bullish here.

I’ve stepped into this layer very aggressively. My partner @anitavlallian and I have acquired almost 6GW coming online in a ramp from today thru 2029 of grid power and behind the meter.

Silicon - I helped get @GroqInc off the ground in 2015 and we licensed it to @nvidia for $20B Dec2025. I won’t invest or incubate anything in this layer now. The perf demands of the chips are too high, manufacturing precision is too complex and supply chain influence to get adjacent components like memory isn’t possible for a startup anymore. Lots of capital will be wasted here chasing Groq and Cerebras’ success. Note that both startups made sense a decade ago when these constraints were much more modest.
Clouds - Clouds are very very lucrative but very hard to build and very expensive and technically complicated to maintain. And as alignment becomes a more important issue, I expect the clouds will be asked to build robust KYC and attest to it. This makes the risk:reward ratio skewed. I don’t want to be responsible when the USG says a cloud allowed a bad actor to do something bad because of poor KYC.
Models are complicated. The big open question is how much of the revenue being generated by them today is because of tokenmaxxing and poor model behavior. If it’s a lot, then the annualized revenues will diminish meaningfully even as token consumption inflects upwards. This is the big economic question at this layer.
Harnesses are where the action is and why I started @8090solutions two years ago. In a nutshell, the harness helps enterprises owns their proprietary context (what Alex Karp calls their ‘alpha’). This is an enterprise’s data, workflows, evals, and business rules. A harness that gives this to an enterprise is what creates very low model-agnostic switching costs, which further reinforces my views of #4 above.
Applications will be another long term winner along with harnesses. This is where the differentiation between “off the shelf” and “custom time and materials” melts away. Every company, with the right harness, can now imbue their alpha into the software that runs their company. I expect this to mean that “off the shelf” is largely replaced with custom software creating a huge opportunity to write these solutions for companies. Build once and sell repeatedly is a laggard GTM motion for a SaaS world that isn’t needed here. Think custom by design, alpha embedded, proprietary by nature.

Fin.

Good luck to all the players!

See 3 related tweets

@omarsar0: recommended reading. the opportunity on harness engineering alone is hard to even measure.

if you ...

@quxiaoyin: I'd add that harness will create margin regardless of if the model gets commoditized. The right harn...
@MilkRoadAI: RT @KyleReidhead: The biggest moat in AI is compute and the hyperscalers are acquiring it FAST

Amaz...

7. _TheTathya (Group Score: 125.1 | Individual: 32.2)

Cluster: 5 tweets | Engagement: 6 (Avg: 51) | Type: Tech

🚨 India’s Energy Security Push Gets Strong Industry Backing

With ₹84,084 crore approved under the Samudra Manthan scheme, India is moving to unlock its vast offshore energy potential.

Pranav Adani’s support highlights an important reality:

Domestic exploration must scale Import dependence must fall Private industry and government must work together Energy security must be built at home

For infrastructure groups like Adani, this opens the door to contribute at scale to India’s next energy chapter.\n\nQT @PranavAdani: India’s energy exploration landscape is entering an era of unprecedented scale. Guided by Hon’ble PM Shri @narendramodi ji’s visionary leadership for India’s energy security, the Union Cabinet’s approval of the Samudra Manthan scheme marks a landmark step towards unlocking our vast offshore energy potential. It will accelerate domestic exploration and production, foster greater industry collaboration and strengthen the foundations of an energy-secure #ViksitBharat 🇮🇳

See 4 related tweets

@Sanjay_Sriv: The Union Cabinet has now formally approved the National Offshore Exploration Scheme, Samudra Mantha...
@mygovindia: Why import energy when the answers are right beneath the ocean?

India just greenlit Samudra Mantha...

@beatsinbrief: 🚨 Adani Group’s Pranav Adani has welcomed the Cabinet’s approval of the ₹84,084 crore Samudra Mantha...
@Sanjay_Sriv: RT @anil7kishan: Samudra Manthan is more than an exploration programme - it is a strategic national ...

8. OrcaRouter (Group Score: 112.0 | Individual: 30.0)

Cluster: 6 tweets | Engagement: 4 (Avg: 11) | Type: Tech

The latest DeepSeek V4 Flash weights appear to represent a genuine capability jump rather than just benchmark optimization.

The architecture doesn't appear to have changed, suggesting the gains come primarily from post-training: stronger RL, higher-quality synthetic coding trajectories, improved tool use, and better long-horizon planning. While some benchmaxxing is inevitable, recent coding benchmarks and agentic evaluations are increasingly difficult to game.

If these gains generalize to real-world repos, Flash approaching frontier proprietary models at ~1/50+th the inference cost would fundamentally shift the economics of coding agents, making intelligent routing more valuable than relying on a single frontier model.

Deepseek V4 Flash 0731 vs Opus 4.8: https://t.co/idDIqzCGcK

DeepSeek V4 Flash 0731 vs GLM 5.2: https://t.co/IMVh6fDQ92\n\nQT @deepseek_ai: 🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!

🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇 🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!

Check out the configuration details in our official API docs: https://t.co/smCwQZMeiq

See 5 related tweets

@VadimStrizheus: IT FINALLY HAPPENED!!

DeepSeek just launched V4-Flash and it’s incredibly shocking..

It is 18x che...

@TeksEdge: What a crazy leap from preview to GA for DeepSeek V4 Flash. But I can’t wait to see what people bui...
@AndrewCurran_: RT @arena: Exciting news: DeepSeek-V4-Flash-High by @deepseek_ai has reshaped the Pareto Frontier in...
@pstAsiatech: Wait, another frontier model from a Chinese AI lab?\n\nQT @poezhao0605: DeepSeek released the offici...
@0x_kaize: RT @0x_kaize: this is INSANE, seriously

DeepSeek just released V4 Flash Official API and the Agent ...

9. mikenevermiss (Group Score: 108.0 | Individual: 36.4)

Cluster: 4 tweets | Engagement: 116 (Avg: 51) | Type: Tech

Anthropic engineer:

“90% of our engineers were building self-improving loops. now we’ve moved to agentic graphs.”

“we’re not prompting anymore.”

In just 10 minutes, she walks through her entire Claude Code workflow live starting from a blank terminal.

this is the kind of knowledge people charge $500+ for.

watch the video first, then save the article below to learn how to become a graph engineer. 👇\n\nQT @mikenevermiss: https://t.co/W4sHoE0ew0

See 3 related tweets

@0xCodez: Anthropic engineer at Stanford:

“80% of our engineers are using self-improving loops. Now everyone ...

@RoundtableSpace: An Anthropic engineer dropped a 2-hour masterclass on graph engineering

“Everyone is now building ...

@0xCodez: RT @0xCodez: Anthropic engineer at Stanford:

“80% of our engineers are using self-improving loops. ...

10. sairahul1 (Group Score: 105.2 | Individual: 33.3)

Cluster: 4 tweets | Engagement: 90 (Avg: 70) | Type: Tech

Mark Cuban on the next job wave.

Customized AI integration for small to mid-sized companies.

"Software is dead because everything's gonna be customized to your unique utilization. Who's gonna do it for them... And there are 33 mn companies in the US."

That person is called a Forward Deployed Engineer.

And it's the hottest job in tech right now and most people have never heard of it.

Here's the no-BS guide to what it is, what it pays, and how to become one ↓

Bookmark this.\n\nQT @sairahul1: https://t.co/p3X0PW3eAZ

See 3 related tweets

@sairahul1: Forward Deployed Engineer is the hottest, and one of the most in-demand, jobs right now.

Every maj...

@sairahul1: RT @sairahul1: Mark Cuban on the next job wave.

Customized AI integration for small to mid-sized co...

@sairahul1: RT @sairahul1: Forward Deployed Engineer is the hottest, and one of the most in-demand, jobs right n...

11. DynamicWebPaige (Group Score: 100.6 | Individual: 35.9)

Cluster: 3 tweets | Engagement: 56 (Avg: 45) | Type: Tech

the most exciting part of this: 800K sign-ups means there are hundreds of thousands of people who want to build software on the go -- most of whom would never call themselves developers!

meeting y'all inside the @GeminiApp, instead of installing yet another tool, is exactly right 🔥\n\nQT @GoogleAIStudio: thank you to the ~800,000 of you that pre-ordered our mobile app on iOS and Android — it’s clear that people are interested in building software on the go

instead of asking you to download yet another app we’ve decided to take an entirely different approach: one where apps emerge naturally, in the course of your everyday conversations with Gemini. we’re partnering with the Gemini app team to make that a reality on mobile and desktop

we’re continuing to invest in the AI Studio web experience for those who want to go from an idea to a prompt to a business

the teams are working hard on bringing this to life, more to share on this later

See 2 related tweets

@IamEmily2050: This is the best way moving forward. I am glad to see the Google AI Studio team and the Gemini app t...
@Gorden_Sun: 谷歌将要整合Gemini APP和Google AI Studio APP，整合的端包括移动端和桌面版，AI Studio的网页版会保留。对谷歌是好事，集中精力把Gemini APP做好。对我是坏...

12. zephyr_z9 (Group Score: 99.3 | Individual: 26.4)

Cluster: 4 tweets | Engagement: 156 (Avg: 386) | Type: Tech

Global R&D spending was $2.82 trillion last year AI4S should be able to capture at least 20% of this spend by 2030\n\nQT @polynoamial: The cost of generating the proofs for all 10 of these breakthroughs combined was under$ 2,000 at Sol API prices. We’re excited to see what scientists and researchers are able to create with our upcoming Astra models!

See 3 related tweets

@firstadopter: Not too shabby. I "Feel the RSI" coming.\n\nQT @polynoamial: The cost of generating the proofs for a...
@willdepue: insane\n\nQT @polynoamial: The cost of generating the proofs for all 10 of these breakthroughs combi...
@tylercowen: RT @polynoamial: The cost of generating the proofs for all 10 of these breakthroughs combined was un...

13. ElliotGlazer (Group Score: 93.8 | Individual: 51.2)

Cluster: 4 tweets | Engagement: 2158 (Avg: 201) | Type: Tech

Oh it leaked. Yeah this is real and probably the most important math AI result yet.\n\nQT @iron_redux: 🚨🚨Breaking news from the math world! Open AI just proved nonsofic groups existed! GPT 5.6 Sol continues to replace mathematicians!!! https://t.co/6xOjD7ZbQM

See 3 related tweets

@alexframegreen: RT @ElliotGlazer: Oh it leaked. Yeah this is real and probably the most important math AI result yet...
@rickasaurus: RT @iron_redux: 🚨🚨Breaking news from the math world! Open AI just proved nonsofic groups existed! GP...
@ElliotGlazer: RT @AcerFur: Very excited about the proven existence of nonsofic groups. To me, this is the first “s...

14. FellMentKE (Group Score: 89.8 | Individual: 34.8)

Cluster: 3 tweets | Engagement: 111 (Avg: 152) | Type: Tech

“Can AI generate a video?” is no longer the interesting question.

The bigger question is whether AI can turn one campaign brief into a consistent set of usable creative variations.

Dreamina Seedance 2.5, with up to 50 multimodal references and 30-second generation, brings that question much closer to reality.\n\nQT @AdarshChetan: For years, AI video has forced creators to choose:

More creative freedom — or more control.

More references — or a manageable workflow.

Longer scenes — or consistent characters, lighting, and story.

That trade-off is exactly what Dreamina Seedance 2.5 is here to challenge.

Dreamina, ByteDance’s proprietary AI product, has unveiled the world-first global launch of Seedance 2.5. Early access is live on Dreamina, while US availability is expected roughly one week later.

With up to 50 multimodal references, 30-second single-generation video, and the lowest-cost access to the full Seedance family, Dreamina is the most cost-effective place to experience Seedance 2.5 at full power.

See 2 related tweets

@heynavtoor: Dreamina Seedance 2.5 has officially launched on Dreamina, the official AI creation platform from By...
@Parul_Gautam7: Dreamina Seedance 2.5 has officially launched on Dreamina, ByteDance's dedicated AI creation hub.

I...

15. swyx (Group Score: 87.0 | Individual: 41.0)

Cluster: 3 tweets | Engagement: 168 (Avg: 57) | Type: Tech

> Google was too nervous to release it and DeepMind was blocked from shipping products that could disrupt Google.

bookmark for the next vc that asks you "what if <incumbent> builds this?"\n\nQT @thsottiaux: @_chenglou I was part of that team. Basically ChatGPT one year before it came out. Called LMChat and then another codename.

Google was too nervous to release it and DeepMind was blocked from shipping products that could disrupt Google.

I think about this a lot.

See 2 related tweets

@SIGKITTEN: one day someone (or something) is gonna make a "the man in the high castle" version of this story\n...
@inafried: RT @thsottiaux: @_chenglou I was part of that team. Basically ChatGPT one year before it came out. C...

16. simonguozirui (Group Score: 86.7 | Individual: 31.7)

Cluster: 4 tweets | Engagement: 200 (Avg: 45) | Type: Tech

RT @miramurati: We share our approach to open-weights releases, how we assessed Inkling, why safety depends on both the model and the ecosystem it enters, and how testing, staged access, and stronger defenses can create a path toward greater openness.

See 3 related tweets

@natolambert: This type of work is incredibly valuable for explaining to the broader public how you can keep relea...
@BlancheMinerva: Really excellent blog post by @thinkymachines, and thrilled to see them putting our work on marginal...
@xeophon: Amazing blog post I very much agree with.

I also think it is more important than ever to enable fr...

17. daleverett (Group Score: 84.4 | Individual: 28.7)

Cluster: 3 tweets | Engagement: 44 (Avg: 14) | Type: Tech

Also we’re looking for a founding engineer.

Since we’re getting so many DMs,

Repost our launch vid with a one liner about yourself to be considered. Or submit a PR on our repo :)\n\nQT @fdotinc: they gave frontier models a 1 billion token context window:

you can now run Fable [1B], Opus [100M], GPT 5.6 [500M] across your entire codebase

it's a graph based tool that can traverse databases with 400M+ of nodes in ms
only one prompt to install
& it's open source (link in video)

Introducing Polygres

See 2 related tweets

@daleverett: All hail the infinite context window 🪟 https://t.co/j7kyxbUbVp\n\nQT @fdotinc: they gave frontier mo...
@daleverett: Holy shit it’s only been a few hours https://t.co/EA86efmM4s\n\nQT @fdotinc: they gave frontier mode...

18. elonmusk (Group Score: 83.3 | Individual: 32.5)

Cluster: 3 tweets | Engagement: 5765 (Avg: 11773) | Type: Tech

Grok Build can do almost anything you can think of https://t.co/lOJK57wDP5\n\nQT @XFreeze: Most people seriously underestimate what Grok Build can do

They assume an AI coding agent is only useful for building apps or writing complex software

But some of its most valuable work is handling the boring, obvious tasks that quietly consume 10 minutes here, 20 minutes there, every single day

For example, I gave Grok Build a SpaceX video posted on 𝕏 and asked it to extract the clearest frames and turn them into a clean collage

It pulled the video, selected the strongest images, cropped them, arranged the layout and produced the final collage shown in the post below

I simply described the result I wanted, then let it handle the entire process while I worked on something else

And that is only one small example

You can use Grok Build to: • Organize and intelligently search files across your computer • Diagnose and fix audio, display and system-setting problems • Download, install and configure apps or tools • Extract images or audio from videos • Add music, trim, compress and convert media • Resize, crop and enhance images • Batch rename and organize hundreds of files • Clean spreadsheets, merge datasets and build dashboards • Write scripts and automate repetitive tasks • Fix broken configurations and connect tools together

Even something as annoying as an audio setting not working becomes much easier

Instead of hunting through menus, searching forums and trying random fixes, you can explain the problem to Grok Build

It can inspect the setup, identify the likely cause, change the relevant configuration and test whether the fix worked

It is basically an agentic operator living inside your computer, working within the permissions you give it

I also keep a simple instruction file telling Grok how I prefer tasks to be completed, including automatically opening the finished result after a successful run so I do not forget about it

Configure the workflow once, and Grok keeps working the way you prefer

That is what makes Grok Build so useful It is not only building software

It turns all the small computer tasks between an idea and the finished result into work you can simply describe and delegate

See 2 related tweets

@XFreeze: Most people seriously underestimate what Grok Build can do

They assume an AI coding agent is only u...

@teslaownersSV: Elon just amplified it: “Grok Build can do almost anything you can think of.” Not just code. File or...

19. coinbureau (Group Score: 79.6 | Individual: 27.7)

Cluster: 3 tweets | Engagement: 188 (Avg: 352) | Type: Tech

JUST IN: Google has scrapped its new Earth AI satellite image generator, just one day after launch, after users flooded the internet with convincing fake satellite imagery.

The feature, integrated directly into Google Earth, allowed realistic AI-generated images of real-world locations, including military sites and disaster zones.

With Google Earth widely relied on by journalists and OSINT investigators as a source of truth, Google has now pulled the feature amid growing concerns over misinformation.

See 2 related tweets

@business: Alphabet’s Google announced Friday that it will roll back its new AI image generation feature in Goo...
@Polymarket: JUST IN: Google rolls back its new Earth AI image generator one day after launch over concerns it co...

20. 0xMovez (Group Score: 78.3 | Individual: 30.6)

Cluster: 3 tweets | Engagement: 95 (Avg: 68) | Type: Tech

Google just released a 1-hour course on “Context Engineering” for AI agents from scratch:

• 00:00 - Introduction to “Context Engineering” • 03:02 - Context Engineering vs “prompting” • 16:18 - Build a context brain for your agent • 22:10 - Building agent context (tools, skills) • 29:06 - Agentic memory bank from scratch

This 1-hour watch could replace a $500 course on agentic design.

Watch it today, then read the full guide on Context Engineering in the article below.\n\nQT @0xRafy: https://t.co/QpI5GJDBmQ

See 2 related tweets

@RoundtableSpace: Google just dropped another amazing video on “Context Engineering”

It covers:

context engineering...
@0xMovez: RT @0xMovez: Google just released a 1-hour course on “Context Engineering” for AI agents from scratc...