科技推特精选 - 2026年4月18日

2026年4月18日科技每日简报

Today's top tech conversations are led by @cryptopunk7213, whose post about 'fucking ruthless lol. now we k...' garnered the highest engagement. Key themes trending across the top stories include design, https, openai, claude, agents. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. cryptopunk7213 (Group Score: 1076.2 | Individual: 70.0)

Cluster: 33 tweets | Engagement: 3974 (Avg: 362) | Type: Tech

fucking ruthless lol. now we know why anthropic left the board of figma this week

they built a product that not only replaces them it’s just better

Figma stock is getting crushed on the news and already down 50% this year 💀

claude design

reads your code base, creates a custom design system for it uses new opus 4.7 to create design assets

this is 90% of the work designers do.\n\nQT @claudeai: Introducing Claude Design by Anthropic Labs: make prototypes, slides, and one-pagers by talking to Claude.

Powered by Claude Opus 4.7, our most capable vision model. Available in research preview on the Pro, Max, Team, and Enterprise plans, rolling out throughout the day. https://t.co/2BgBGtgYGX

See 32 related tweets

@Yuchenj_UW: Figma stock 20 minutes after the Claude Design announcement.

Wild. https://t.co/oe2vZl8Q91\n\nQT @c...

@AndrewCurran_: RT @claudeai: Introducing Claude Design by Anthropic Labs: make prototypes, slides, and one-pagers b...
@Hesamation: watching Figma, a $10B UX design monopoly, melt down with a single Claude tweet. https://t.co/7OJRaV...
@wallstengine: Anthropic Labs launched Claude Design, a new product for creating visual assets, prototypes, slides,...
@JoshKale: Three days ago, Anthropic's CPO left Figma's board. Today, Anthropic became Figma...

Here's everyth...

2. testingcatalog (Group Score: 185.1 | Individual: 33.3)

Cluster: 7 tweets | Engagement: 364 (Avg: 278) | Type: Tech

Anthropic released Claude Design 👀

A new tool to generate UI prototypes, slides, and one-pagers. Claude Design will be gradually rolled out to all paid plans shortly. https://t.co/YzG6CwO32Z\n\nQT @claudeai: Describe what you want, and Claude builds the first version. Refine through conversation, inline comments, direct edits, or custom sliders.

Export to @canva, as PDF or PPTX, or hand off to Claude Code when the design feels right. https://t.co/EGZ2UmIm6m

See 6 related tweets

@DataChaz: design SaaS right now https://t.co/6MAM9skBqb\n\nQT @claudeai: Introducing Claude Design by Anthropi...
@tunguz: ship shippety ship\n\nQT @claudeai: Introducing Claude Design by Anthropic Labs: make prototypes, sl...
@StockSavvyShay: Anthropic just introduced Claude Design which is a tool that lets users create prototypes, slides &a...
@TechCrunch: Anthropic launches Claude Design, a new product for creating quick visuals https://t.co/AgCEEyNsoF...
@heyshrutimishra: BREAKING: ANTHROPIC JUST LAUNCHED CLAUDE DESIGN.

Auto-applies your design system to every project...

3. aakashgupta (Group Score: 160.3 | Individual: 35.4)

Cluster: 5 tweets | Engagement: 17 (Avg: 119) | Type: Tech

OpenAI's stock-based comp averaged $1.5M per employee last year. That's across 4,000 people. Senior AI PM packages there clear$ 800K total comp. Staff level crosses $1M with equity.

Run the math on what a single interview rejection costs. A candidate who walks into an AI system design round and spends 20 minutes on user personas just left $3M+ of 4-year equity on the table. Then they go home and open another product sense prep doc.

The bid for AI PM talent has disconnected from what 95% of PM candidates are training for. Meta pulled senior AI researchers last year with individual packages that reportedly crossed $100M. OpenAI removed its 6-month vesting cliff to keep people. Anthropic and Google are paying to match.

The ones who get through are not studying harder. They're studying different things. Aakash just documented exactly what.\n\nQT @aakashgupta: Aman just landed an AI PM role.

Here's what he didn't face once in the loop: product design. "Design a pencil for the blind." "Create a dating app for Facebook." The classic PM interview staples every candidate drills for never showed up.

One round kept appearing instead: AI system design.

Data pipelines. Model selection trade-offs. Orchestration layers. Agent architectures. Failure modes. Draw it on a whiteboard, explain why XGBoost beats GPT-4 for churn prediction, and surface the failure modes before the interviewer asks.

This round didn't exist 18 months ago.

Now it's at OpenAI, Anthropic, Google DeepMind, Meta AI, and Amazon AGI. The stakes are absurd. OpenAI stock-based comp averages $1.5M per employee per year. Senior AI PM total comp at these companies clears$ 500K to $800K. Staff-level crosses$ 1M with equity.

Here's the gap that blows my mind. Hundreds of guides on product sense interviews. Hundreds on behavioral. Zero on AI system design for PMs.

I've coached 200+ candidates through AI PM loops in the last 12 months and watched smart people get rejected because they ran the product design playbook. Twenty minutes on user personas. A vague "AI layer" sketched at minute 35. Rejection email by Friday.

The ratio has to flip. 30-40% product framing. 60-70% system architecture. Named agents with non-overlapping responsibilities. Specific models with specific reasoning (XGBoost for prediction because it's cheaper, faster, interpretable; LLM for the voice agent because you need natural language). Failure modes surfaced proactively. A metric cascade that connects model performance to business impact.

So I built the guide. The DASME framework with architecture diagrams. Company-by-company breakdowns for OpenAI, Anthropic, DeepMind, Meta, and Amazon. 7 anti-patterns with the exact fix for each. 64 practice questions. 2 fully worked examples (enterprise RAG + content moderation). Three practice tools (Claude Skill, GPT, Gem) you can mock interview with right now.

Full guide: https://t.co/yTYvPBjTP8

See 4 related tweets

@aakashgupta: Smart PMs are getting rejected from AI loops because of one specific reflex: reaching for the LLM on...
@aakashgupta: OpenAI, Anthropic, and Google DeepMind all quietly killed the product design interview round in the ...
@aakashgupta: 90% of AI PM interview prep is aimed at case interviews. Case rounds show up in roughly 10% of actua...
@aakashgupta: The interview question gating $1M+ AI PM roles at OpenAI, Google, and Meta: "build a churn reduction...

4. Parul_Gautam7 (Group Score: 157.1 | Individual: 29.0)

Cluster: 6 tweets | Engagement: 77 (Avg: 80) | Type: Tech

Editing just got a whole lot smarter — fix the moment, not the whole clip 🔥

Because every second matters, and perfection shouldn’t mean starting over.\n\nQT @Buzzy_now_AI: Introducing the World's #1 Vibe Video Photoshop!

Say goodbye to AE, PR and Capcut. Now you can just chat to edit your videos using Buzzy.

Unlike other video generators, Buzzy fix the exact part without touching the rest of your clip.

Stop reshooting. Start Video "Photoshop".

RT + Comment "Buzzy NOW" to get free 2000

See 5 related tweets

@ai_explorer25: Your video looks almost perfect then one small detail throws it off

That last bit is where things u...

@Origin_AI_01: Chat + video editing doesn’t feel like editing anymore It feels like directing outcomes instead of f...
@wallstengine: Buzzy launched an AI chat-based video editor.

Instead of regenerating the whole video, it lets user...

@TheoBuildsAI: We’ve been re-recording entire clips for one small mistake. This completely kills that workflow. htt...
@Origin_AI_01: RT @Buzzy_now_AI: Introducing the World's #1 Vibe Video Photoshop!

Say goodbye to AE, PR and Capcut...

5. levie (Group Score: 145.1 | Individual: 38.8)

Cluster: 4 tweets | Engagement: 1581 (Avg: 1742) | Type: Tech

Agents are going to use software 100X more than people will in the future. As a result, enterprise platforms will become headless and be able to work with any agent on or off platform. If you don’t do that you’re DOA.

What some have missed is that this creates vastly more use-cases for these platforms than even existed pre-AI. This isn’t zero sum. Software value props have traditionally been capped at the number of users you have in a company. Agents have no upper limit.

We’re going to run agents to process data at a scale humans never could, they’re going to be running 24/7 in parallel doing work for us, and they can integrate workflows across systems to generate all new value propositions.

Once you embrace this approach, it becomes obvious how much more upside there is.\n\nQT @Benioff: Welcome Salesforce Headless 360: No Browser Required! Our API is the UI. Entire Salesforce & Agentforce & Slack platforms are now exposed as APIs, MCP, & CLI. All AI agents can access data, workflows, and tasks directly in Slack, Voice, or anywhere else with Salesforce Headless 360. Faster builds, agentic everything. 🚀 #Salesforce #Agentforce #AI https://t.co/mxySdJS7HR

See 3 related tweets

@rseroter: RT @levie: Agents are going to use software 100X more than people will in the future. As a result, e...
@garrytan: The future is MCP/CLI\n\nQT @Benioff: Welcome Salesforce Headless 360: No Browser Required! Our API...
@nummanali: Industry is moving in the right direction

Headless SaaS is the way forward

Build with an API firs...

6. aakashgupta (Group Score: 138.6 | Individual: 51.9)

Cluster: 4 tweets | Engagement: 717 (Avg: 119) | Type: Tech

OpenAI quietly solved the single problem that's been blocking computer-use agents from going mainstream: the cursor war.

Anthropic shipped computer use in October 2024. OpenAI shipped Operator. Google shipped Mariner. All three hit the same wall. When the agent moves the cursor, you can't. The screen is one resource, and two drivers cannot share it. Every demo ended the same way: "run this while I go get coffee."

That kills the ROI math. If the human has to stop working to let the agent work, the agent is functionally a replacement. Adoption of computer-use agents flatlined at power users and scripted demos for exactly this reason.

Here's what OpenAI actually shipped. Codex runs multiple agents against your real Mac apps, in parallel, with their own cursors, in the background, while you keep using your machine. The mechanism is OS-level sandboxing. OpenAI acquired Sky Applications last fall. That's the team that built Workflow, which became Apple Shortcuts. They spent a decade figuring out how to let third-party code drive other Mac apps without breaking the foreground experience. OpenAI didn't buy a team. They bought the only people on Earth who had solved this.

The economic unit of white-collar work just changed. Yesterday the ceiling was one human with one agent doing one task. Today it's one human supervising five agents working across five apps the human is simultaneously using. The bottleneck stops being compute. It becomes the human's ability to review output.

Codex has 3M weekly devs. The feature that matters is not the IDE. It's that OpenAI turned the Mac into a multiplayer instrument.\n\nQT @sama: Lots of major improvements to Codex!

Computer use is a real update for me; it feels even more useful than I expected. It can use all of the apps on your Mac, in parallel and without interfering with your direct work.

See 3 related tweets

@aakashgupta: Perplexity and OpenAI both shipped major computer-use agents today. Anthropic shipped Cowork earlier...
@romainhuet: RT @aakashgupta: OpenAI quietly solved the single problem that's been blocking computer-use agents f...
@rohanpaul_ai: RT @rohanpaul_ai: OpenAI just expanded Codex from a coding assistant into a desktop agent that can s...

7. yacineMTB (Group Score: 127.3 | Individual: 49.2)

Cluster: 4 tweets | Engagement: 2962 (Avg: 267) | Type: Tech

RT @Alibaba_Qwen: ⚡ Meet Qwen3.6-35B-A3B：Now Open-Source！🚀🚀

A sparse MoE model, 35B total params, 3B active. Apache 2.0 license.

🔥 Agentic coding on par with models 10x its active size 📷 Strong multimodal perception and reasoning ability 🧠 Multimodal thinking + non-thinking modes

Efficient. Powerful. Versatile. Try it now👇

Blog：https://t.co/EXx5y466su Qwen Studio：https://t.co/bg4tAU1p74 HuggingFace：https://t.co/w4pDX14DZS ModelScope：https://t.co/SuRyLzdQiO API（‘Qwen3.6-Flash’ on Model Studio）：Coming soon～ Stay tuned

See 3 related tweets

@victormustar: Sharing my current setup to run Qwen3.6 locally in a good agentic setup (Pi + llama.cpp). Should giv...
@WesRoth: Alibaba open-sourced Qwen3.6-35B-A3B under a permissive Apache 2.0 license.

It is a highly efficie...

@QuixiAI: This happened because of Gemma 4\n\nQT @Alibaba_Qwen: ⚡ Meet Qwen3.6-35B-A3B：Now Open-Source！🚀🚀

A s...

8. chddaniel (Group Score: 124.2 | Individual: 32.2)

Cluster: 4 tweets | Engagement: 8 (Avg: 29) | Type: Tech

Today, Shipper built everything Lovable was meant to be.

Shipper is now able to build any: → Mobile app → Website/web app → Chrome browser extension → Bot for Telegram, Discord, Slack, WhatsApp → Convert website to mobile app and vice-versa

Underneath, CC w/ Opus 4.7 is the brain. It breaks your prompt into subtasks and routes each one to whichever of 20 models wins at it. GPT for long context. Opus for deep research. Veo for motion design. Remotion for video.

It runs 24/7. You can run it on your phone or computer.

This is, frankly, insane.\n\nQT @shipper_now: Introducing the First AI Revenue Agent that vibe codes for you.

Build AI that does deep research, codes, designs, monetizes, deploys, translates, and writes email sequences.

One-shots Claude, Duolingo, Fiverr, and more.

To celebrate the launch, we're giving away free credits randomly. Repost and reply "Ship" below and let Siri pick the winners 💥

See 3 related tweets

@chhddavid: Today Shipper shipped everything Lovable was meant to be 🚢

I just watched my Mac oneshot any form o...

@chddaniel: so you're saying that Claude Opus 4.7
oneshots any billion-dollar company
launches it form
ru...
@chddaniel: This is terrifying...\n\nQT @shipper_now: Introducing the First AI Revenue Agent that vibe codes for...

9. niccruzpatane (Group Score: 115.6 | Individual: 50.0)

Cluster: 4 tweets | Engagement: 24844 (Avg: 596) | Type: Tech

RT @elonmusk: Universal HIGH INCOME via checks issued by the Federal government is the best way to deal with unemployment caused by AI.

AI/robotics will produce goods & services far in excess of the increase in the money supply, so there will not be inflation.

See 3 related tweets

@Cointelegraph: ⚡️ NOW: Elon Musk says Universal High Income via government checks is the best solution to AI-driven...
@MyLordBebo: The likely scenario is that, if robots will run the economy and many humans are useless for this pro...
@Rasmic: All the permanent underclass stuff sounded like a joke until I read this\n\nQT @elonmusk: Universal ...

10. rohanpaul_ai (Group Score: 111.4 | Individual: 50.3)

Cluster: 3 tweets | Engagement: 720 (Avg: 48) | Type: Tech

RT @TencentHunyuan: We’re open-sourcing HY-World 2.0, a multimodal world model that generates, reconstructs, and simulates interactive 3D worlds from text, images, and videos.

Outputs can be integrated into game engines and embodied simulation pipelines.

Key highlights: 🔹 One-click world generation Turn text or image into interactive 3D worlds automatically. 🔹 Pipeline-ready 3D outputs Editable 3D worlds for Unity and Unreal Engine, with standard 3D exports including mesh, 3DGS, and point clouds. 🔹 Unified world model system One model family for world generation and reconstruction across synthetic and real-world scenes. 🔹 Interactive character mode Explore generated 3D worlds in real time with physics-aware movement and collision support.

✨ Apply for access: https://t.co/swscD5KGu2 🔗 GitHub: https://t.co/XpUKjBtK5n 🤗 Hugging Face: https://t.co/tv8hOPYABj 📄 Technical Report: https://t.co/s6WGMyw0L7

See 2 related tweets

@WesRoth: Tencent open-sourced HY-World 2.0, an advanced multimodal world model engineered to generate, recons...
@rohanpaul_ai: Really brilliant release from Tencent.

HY-World 2.0, an open 3D world model that turns text, image...

11. chatgpt21 (Group Score: 102.1 | Individual: 28.2)

Cluster: 4 tweets | Engagement: 62 (Avg: 142) | Type: Tech

The keyboard is 150 years old and we still act like it is the final form of computing.

If Sabi is right, the next big interface is not another screen. It is non-invasive BCI good enough that you can just think and the system responds.

Maybe you seen comments with me saying that the keyboard will be replaced by voice, but this is a much better idea\n\nQT @rahulchhabra07: you can now control things with your brain. literally.

we're building the most wearable BCI on the planet, with @sabicap, backed by @khoslaventures @accel @initialized & @kevinweil.

we collected the world’s largest neural dataset and trained the most capable Brain Foundation Model.

then we invented a new class of biosensors powered by custom ASICs.

type without typing. click without clicking. a cap that lets your brain do the work.

we’re sabi.

See 3 related tweets

@rohanpaul_ai: RT @rohanpaul_ai: This is the AlphaFold moment for brain-computer interfaces.

Sabi releases the Sa...

@jack: RT @rahulchhabra07: you can now control things with your brain. literally.

we're building the most ...

@ArjunShukl95550: what if baldness is just an adaptation to enable better brain<> computer connection?

i mean, ...

12. aakashgupta (Group Score: 100.1 | Individual: 35.3)

Cluster: 3 tweets | Engagement: 15 (Avg: 119) | Type: Tech

The fastest adopters of AI design tools have zero design background.

GTM people and finance operators with no design training. They opened the tools, started prototyping, and realized they could produce in an afternoon what used to require a two-week design sprint. No formal adoption roadmap. They just started.

Meanwhile, large design orgs at companies like UnitedHealth and Ginkgo Bioworks are still debating access policies.

OpenAI's Ed Bayes and Figma's Gui Seiz walked through how this plays out in a live demo. Code-based generation on one side, canvas-based on the other, Figma MCP connecting them. The skill floor for building a product interface dropped to near zero while the large teams were still planning rollouts.

Here's why this pattern keeps repeating. The people with the least process to protect always move first. Designers have design systems and review workflows built over years. That infrastructure is valuable. It also creates real friction against adopting a tool that bypasses half of it. The GTM person prototyping a UI has nothing to unlearn. They just build.

The implication for PMs is direct. When anyone on the team can go from idea to interactive prototype in one sitting, the handoff chain from spec to design to review compresses into one workflow. The PM who prototypes their own concept stops waiting in a queue.

The roadmap is the first session.\n\nQT @aakashgupta: This is how OpenAI and Figma design.

I got a masterclass from @OpenAI's @edbayes and @figma's @guiseiz:

1:33 - Code vs canvas 6:00 - When to use which 10:06 - Live demo 13:01 - Figma MCP 33:13 - Adoption roadmap 43:29 - Roles blurring https://t.co/Ybx4wzRhRj

See 2 related tweets

@aakashgupta: The PM role most people are preparing for doesn't exist at the companies building the future.

At Op...

@aakashgupta: The design process hasn't changed in a decade. That's exactly why most designers will miss what's ha...

13. teortaxesTex (Group Score: 97.3 | Individual: 35.5)

Cluster: 3 tweets | Engagement: 76 (Avg: 54) | Type: Tech

my low effort steelman of Jensen's holistic position on why he should be allowed to sell chips to China, with a strategic framing conclusion. For those of you who prefer "rational" loser talk to a hyper-successful CEO's grug-brained noises. https://t.co/L5ZmJ6IIdM\n\nQT @teortaxesTex: I think this is the usual problem with rationalists. You overestimate how well you get the topic, due to the skill in making rhetorically defensible arguments, and underestimate how much domain knowledge is needed to make an object-level correct prediction. In this case the most important facts are about dynamics of PRC hardware development, and how they change conditional on American presence in the market. Politically, rationalists tend to be smugly naive and autistic, more brainwashed and hawkish than a typical neocon, so default to bullshit like "Communist Chyna will fund Huawei to parity anyway irrespective of profitability". That's not what has been happening historically. The scale of necessary subsidies is beyond even Chinese state's pain threshold, Huawei doesn't get infinite subsidies, and cannot develop an ecosystem without actual power users, so they've been stagnating while Nvidia hardware was abundant. China can't just mandate that power users move to Huawei (or Cambricon or anything) from available Nvidia hardware either – not without disrupting their companies and having massive brain drain. But over the period of externally imposed controls, we have seen decent progress in systems-level design and adoption. They're now near the threshold where the domestic stack is just usable. They have serious companies like ZAI adopting it, a growing volume of publications, improving software support, growing utilization of networking advantages of HW. It's still garbage compared to Nvidia/CUDA, but not unusable garbage; just finicky and Pareto-inferior. Another year of this, and it'll likely achieve criticality – they'll have massive profits, tons of feedback, and begin iterating at Nvidia's pace. This will erode the factor of vendor lock-in that currently exists wrt CUDA. After that, Jensen will have an exponentially harder job every year getting back into that market and taking any piece of it.

In the long run, the result is just that the US loses the biggest slice (China) of one of the world's largest markets (compute substrate), and is that much more diminished and irrelevant. This is the modal, default outcome: China becomes the biggest economic entity in the world, the central node, and tapping out of any Chinese market where you had legitimate technological advantage today is shooting oneself in the foot tomorrow. Moreover, doing this with the intent to choke, dominate and possibly regime change them could invite retaliation (at least I'd certainly seek to defang such an unhinged and paranoid state after gaining the upper hand; who knows what terrorism it'd stoop to if given an occasion). I understand that rationalists, patriots, other hawks assume that the US can, by default, not only prevent Chinese economic primacy, but establish some durable strategic advantage (via AGI, recursive self-improvement etc., plus Starship and other cards) and make China irrelevant; and that this opportunity may be lost if we're selling them compute. I believe it is irresponsible to proceed from this assumption. But this depends on all sectors of one's payoff matrix.

See 2 related tweets

@teortaxesTex: «hey ASML, stop servicing old DUV machines» is pretty much the last rung on the export control escal...
@teortaxesTex: Jensen is a Taiwanese. He only learned Mandarin while working at AMD. I find it odd how people assum...

14. burkov (Group Score: 92.0 | Individual: 63.6)

Cluster: 2 tweets | Engagement: 8008 (Avg: 264) | Type: Tech

For those living under a rock: LLMs stopped becoming smarter around summer 2025.

Everything impressive you see since then is about finetuning them for specific tasks (mainly coding and software-tool-based task solving) and building tooling around them (such as agentic coding systems).\n\nQT @minchoi: Claude Opus 4.7 has achieved AGI https://t.co/hAtdkZComH

See 1 related tweets

@burkov: RT @burkov: For those living under a rock: LLMs stopped becoming smarter around summer 2025.

Everyt...

15. ThisWeeknAI (Group Score: 91.1 | Individual: 32.0)

Cluster: 5 tweets | Engagement: 80 (Avg: 64) | Type: Tech

One shotted by Opus 4.7

The full prompt: "Make a website for This Week in AI podcast with great branding." https://t.co/rbHP2qTyQB\n\nQT @ThisWeeknAI: Opus 4.6 vs. Opus 4.7 for website design:

4.6 looks like AI slop. 4.7 looks like it was designed by a $20k team.

The leap in quality is insane. https://t.co/Ql35GdEoqL

See 4 related tweets

@ThisWeeknAI: The prompt: "Make a website for This Week in AI podcast with great branding."\n\nQT @ThisWeeknAI: Op...
@Shipper_now: oh wow... Claude Opus 4.6 is so good at design!\n\nQT @chhddavid: 🚨 Opus 4.6 vs. Opus 4.7 for web de...
@jason: RT @ThisWeeknAI: Opus 4.6 vs. Opus 4.7 for website design:

4.6 looks like AI slop. 4.7 looks like ...

@RoundtableSpace: The difference in quality in web design from Opus 4.6 to 4.7 is insane https://t.co/D4znnul05q...

16. svpino (Group Score: 88.8 | Individual: 33.4)

Cluster: 3 tweets | Engagement: 46 (Avg: 157) | Type: Tech

Back in the early 2000s, I made a living turning Photoshop designs into working web apps.

I made a ton of money doing this.

You can now do the same with two clicks.

We may not think about this anymore, but there were tens of thousands of people who made a living just like I did.

Not anymore.

We all had to find a different path.

Not the first time this happened to me, and it wasn't the last one.\n\nQT @Base44: Turn your Figma design into a working app in minutes. Just paste your @figma frame link, hit generate, and it's live. https://t.co/nVmTJT1EId

See 2 related tweets

@Parul_Gautam7: Base44 just raised the bar!

• Figma → App in minutes • Built-in Themes • Asset Library • Full Visua...

@TheoBuildsAI: Paste a Figma link Get a working app Real logic behind every screen

This is how it should have work...

17. etnshow (Group Score: 88.3 | Individual: 33.6)

Cluster: 4 tweets | Engagement: 32 (Avg: 20) | Type: Tech

FULL INTERVIEW: JAMES WISE - CHAIR OF SOVEREIGN AI - NEW £500M UK AI FUND

James Wise (@Jameswise) Chairman of Sovereign AI (@UKSovereignAI) & Partner at Balderton Capital (@balderton) joins us live to break down the exact plan for the UK’s £500 million bet to back homegrown AI founders, drive growth and create jobs across the UK:

(00:23) Overview of the Sovereign AI Fund (01:58) Founder Experience and Support Levers (03:50) Integration with Government (06:19) Building Frontier Labs in the UK (08:05) Is £500M Enough? (10:11) State of Compute Infrastructure in UK (11:52) Can The UK Best Anthropic & OpenAI (14:14) All Birds Pivots to AI\n\nQT @UKSovereignAI: Sovereign AI. Launching Tonight. 🇬🇧 https://t.co/8SnHJ8biQs

See 3 related tweets

@etnshow: "UK AI companies in the last quarter raised six billion pounds...3x what we did last year, more than...
@etnshow: RT @etnshow: "It's a £500 million commitment from the British Government to invest like a venture in...
@matthewclifford: RT @KanishkaNarayan: We are launching @UKSovereignAI £500m today for one reason: British AI capabili...

18. tunguz (Group Score: 88.2 | Individual: 58.0)

Cluster: 2 tweets | Engagement: 3184 (Avg: 76) | Type: Tech

RT @HeyGen: We built our launch video in Claude Code using HyperFrames.

Now it's yours.

Open source, agent-native framework. HTML to MP4.

$ npx skills add heygen-com/hyperframes

RT + Comment "HyperFrames" to get the full source code of this launch video (must follow) https://t.co/vsRtZ6gQsb

See 1 related tweets

@wshuyi: 用 Claude Code 调用 HyperFrames Skill 试了一下，主题为「深度神经网络分类结构化数据」，效果很炫。但实话实说，在我的机器学习教学场景下，没有发现比起 Remotion V...

19. googledevs (Group Score: 85.8 | Individual: 40.4)

Cluster: 3 tweets | Engagement: 430 (Avg: 120) | Type: Tech

A2UI v0.9 is the new standard for portable Generative UI, allowing your agents to natively "speak" UI directly to your existing frontends.

✨ React, Flutter, and Angular support 🐍 Python SDK via pip install ⚡ High-performance streaming 🛠️ Design system integration

Learn more via the blog: https://t.co/YOTjKGBDjm

See 2 related tweets

@rseroter: It's getting easier for agents to generate UIs when they need to interact with people.

A2UI v0.9 is...

@CopilotKit: RT @ataiiam: Today, we're joining Google's release of A2UI v0.9 🎉

Declarative Generative UI with yo...

20. BrianRoemmele (Group Score: 83.3 | Individual: 32.6)

Cluster: 3 tweets | Engagement: 70 (Avg: 496) | Type: Tech

Time Prices Are The Only Valid Human Metric.

The Simon Abundance Index obliterates the shortcomings if how economies really work by using time prices as its foundational unit: the number of hours an average worker must labor to earn enough to buy one unit of a commodity, calculated simply as nominal market price divided by hourly after tax wage.

Time prices are universal, inflation proof, and deeply personal. They translate abundance into the currency that matters most, hours of your life. When time prices fall, you literally get more for less of your finite time on Earth. Consider a pound of coffee.\n\nQT @BrianRoemmele: KRYPTONITE TO DOOMER

The Simon Abundance Index and the broader Simon Abundance Framework, provides the rigorous, data driven talisman we will wield today.

Simon’s legacy is not merely academic. It is profoundly philosophical and practical. He showed that optimism grounded in evidence is not naive.

It is the only rational stance. In the Age of AI and robotics his vision has never been more relevant.

Every new human life, empowered by exponentially growing technology, multiplies our capacity to solve problems and create value at unprecedented scales.

See 2 related tweets

@BrianRoemmele: KRYPTONITE TO DOOMER

The Simon Abundance Index and the broader Simon Abundance Framework, provides...

@BrianRoemmele: RT @BrianRoemmele: KRYPTONITE TO DOOMER

The Simon Abundance Index and the broader Simon Abundance ...