SoftBank is in talk with Nvidia and Foxconn to build domestically produced AI servers as part of Japan’s AI infrastructure push
*Walter Bloomberg @DeItaone
$NVDA - SOFTBANK IN TALKS WITH NVIDIA TO BUILD MADE-IN-JAPAN AI SERVERS - NIKKEI
5:05 PM · May 7, 2026 · 106K Views
45 Replies · 70 Reposts · 773 Likes
Crypto Integrated is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Cloudflare is cutting 1,100+ jobs, ~20% of its workforce and shifting to an AI-first model
Cointelegraph @Cointelegraph
🚨 NOW: Cloudflare plans to slash 1,100+ jobs, making up one-fifth of its workforce as it shifts to an AI-first operating model.
3:20 AM · May 8, 2026 · 11.7K Views
33 Replies · 14 Reposts · 74 Likes
Coinbase’s x402 and wallet infrastructure are now natively integrated into Amazon Bedrock AgentCore Payments for USDC micropayments on Base and Solana
Brian Armstrong @brian_armstrong
Now AI agents on AWS can pay for services in USDC, settled on
3:48 PM · May 7, 2026 · 381K Views
252 Replies · 347 Reposts · 2.77K Likes
AWS Bedrock AgentCore Payments also integrated Privy wallet infrastructure for agent spending controls and Stripe onramp funding
1/ Today, we're proud to announce that Bedrock AgentCore payments is integrating Privy wallet infrastructure.
This helps developers develop agentic payment solutions that can reason, act, and now transact—with payments built directly into how they operate.
3:29 PM · May 7, 2026 · 19.2K Views
8 Replies · 15 Reposts · 158 Likes
Tether AI released QVAC MedPsy, an open-source on-device medical AI, the 4B model outperforms MedGemma 27B on clinical benchmarks
Paolo Ardoino 🤖 @paoloardoino
We just released our QVAC MedPsy, Tether AI SoTA medical health AI model, capable of high-performance execution and high-accuracy directly on smartphones, laptops and servers.
Highlights:
- QVAC MedPsy 4B beats MedGemma 27B
- QVAC MedPsy 1.7 beats MedGemma 4B
- 3.2x reduction in
8 billion humans deserve an intelligence that doesn't blink when the signal dies. 🧠
Introducing @QVAC Psy, our foundational models built on the mathematical stability of Psychohistory.
With QVAC MedPsy, our local-first medical health AI model, we’ve proven that superior
12:18 PM · May 7, 2026 · 21.5K Views
18 Replies · 20 Reposts · 171 Likes
ElevenLabs launched Studio Agent inside ElevenCreative, an AI co-editor that builds video drafts on the timeline from a text prompt, places sound effects frame-accurately, and lets users interrupt and hand control back at any point
Introducing Studio Agent in ElevenCreative.
Studio is the timeline editor where creators and marketers mix voiceovers, music, sound effects, and video into finished content. Now it has an AI co-editor built in.
5:00 PM · May 7, 2026 · 826K Views
11 Replies · 23 Reposts · 218 Likes
Anthropic donated its open-source alignment tool Petri to Meridian Labs and released Petri 3.0, featuring major upgrades in adaptability, realism, and depth, including a new Dish add-on to make evaluations harder to detect
We’re donating Petri, our open-source alignment tool, to , so its development can continue independently.
Working with Meridian Labs, we’ve also released a major update that improves the adaptability, realism, and depth of Petri’s tests.
anthropic.com
Donating our open-source alignment tool
9:03 PM · May 7, 2026 · 77.5K Views
58 Replies · 73 Reposts · 965 Likes
Anthropic’s security bug bounty program is now publicly available on HackerOne, offering rewards between $100 and $10,000
Our security bug bounty program is now public on HackerOne.
We've run the program privately within the security research community, and their findings have strengthened our products. Now anyone can report vulnerabilities and get rewarded.
Read more:
hackerone.com
Anthropic - Bug Bounty Program | HackerOne
7:10 PM · May 7, 2026 · 167K Views
104 Replies · 191 Reposts · 1.88K Likes
OpenAI launched three new API audio models: GPT-Realtime-2 for voice reasoning, GPT-Realtime-Translate for live speech translation across 70+ languages, and GPT-Realtime-Whisper for live transcription
Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents.
Voice agents are now real-time collaborators that can listen, reason, and solve complex problems as conversations unfold.
Now available in the API
5:19 PM · May 7, 2026 · 756K Views
387 Replies · 599 Reposts · 7.45K Likes
OpenAI’s Codex Chrome extension is officially rolling out on macOS and Windows, enabling parallel background browser work across tabs without taking over the browser
OpenAI Developers @OpenAIDevs
Codex can now take on more of your browser dev work.
With the new Chrome plugin in the Codex app, it can test web apps, gather context across tabs, use web DevTools efficiently in parallel, and keep results organized without taking over your browser.
8:10 PM · May 7, 2026 · 138K Views
72 Replies · 115 Reposts · 1.78K Likes
CoreWeave posted Q1 revenue of $2.08B (up 112% YoY), secured $40B in new customer commitments, surpassed 1GW of active power, and reported a backlog of $99.4B
Q1 2026 results are in:
✅ $40B in new customer commitments
✅ Revenue reached $2.1B, up 112% yoy
✅ Surpassed 1GW of active power
Thank you to our customers for a great start to FY26.
Get the full details:
$CRWV
8:21 PM · May 7, 2026 · 40.5K Views
15 Replies · 36 Reposts · 260 Likes
Signs of a Grok Computer feature appeared in the Grok web UI, with a new selector letting users choose between Grok Computer and a Google Drive folder
Grok computer is here?!
4:55 PM · May 7, 2026 · 18.4K Views
1 Repost · 18 Likes
Google’s Gemini 3.1 Flash Lite is now generally available, optimized for high-volume agentic tasks, translation, and cost-efficiency
Google AI Studio @GoogleAIStudio
gemini 3.1 flash-lite is here
it's our most cost-efficient model, optimized for high-volume agentic tasks, translation, and simple data processing
6:21 PM · May 7, 2026 · 333K Views
147 Replies · 279 Reposts · 2.97K Likes
Meta is internally testing an always-on AI agent called Hatch, grounded in Instagram and Facebook data, with a mock environment rollout targeting end of June
TestingCatalog News 🗞 @testingcatalog
META 🚨: An upcoming always-on AI Agent called "Hatch" from Meta will be available on a waitlist and grounded in social data from Instagram and Facebook.
> According to The Information, Meta is targeting internal testing of Hatch by the end of June, with mock environments built
2:40 PM · May 7, 2026 · 22.6K Views
12 Replies · 13 Reposts · 247 Likes
Scale AI launched the SWE Atlas Refactoring Leaderboard, a new benchmark requiring 2x the code changes of SWE-Bench Pro; Claude Code with Opus 4.7 tops the rankings, followed by Codex with GPT-5.5
Today we’re releasing Refactoring, the final leaderboard of our SWE Atlas suite.
This new leaderboard is the ultimate test of an agent's ability to restructure code without breaking the system.
Claude Opus 4.7 with Claude Code takes the top spot🥇
5:04 PM · May 7, 2026 · 61.5K Views
27 Replies · 37 Reposts · 515 Likes
Kite AI partnered with Ankr to provide the RPC infrastructure layer for its AI agent payment L1
KITE AI 📍 Consensus Miami 2026 @GoKiteAI
Thrilled to partner with on launching Kite RPCs across their global infrastructure!
With Ankr’s reliable, low-latency RPC layer, we're making it easier than ever for developers to build at scale in the agentic economy.
Our secure low-fee micropayments and verifiable
Ankr has partnered with @GoKiteAI ✨
Kite is the first L1 built for AI agents to pay and transact autonomously. Ankr will provide the RPC layer to keep it reliable at scale.
Who benefits?
Every developer building for the agentic economy!
2:46 PM · May 7, 2026 · 4.45K Views
6 Replies · 13 Reposts · 75 Likes
Anthropic is testing an Insights feature on Claude Console that scores agent sessions 0–100 and surfaces cross-session error patterns and usage trends
TestingCatalog News 🗞 @testingcatalog
Anthropic is testing the Insights feature for its Managed Agents on Claude Console.
> Up to 100 recent sessions are fetched. Each transcript is sent to the model (4 in parallel) with your agent's system prompt as context. The model writes a summary — task, actions, issues,
11:49 AM · May 7, 2026 · 8.84K Views
5 Replies · 13 Reposts · 152 Likes
Google is preparing an Agent Mode on Flow for end-to-end video production, including scene planning, generation, and project management from chat
TestingCatalog News 🗞 @testingcatalog
GOOGLE 🚨: A new Agent Mode is being prepared on Flow to handle complete video production process!
Users will be able to:
- Plan out scenes
- Discuss in-progress project changes
- Trigger generation workflows
- Manage both project-level and app-level creative tools
- Update
2:20 PM · May 7, 2026 · 18.6K Views
17 Replies · 28 Reposts · 404 Likes
StrikeRobot partnered with Venice AI as its primary VLM inference backend for SR Agentic and SR Platform
Strike Robot @StrikeRobot_ai
IT'S OFFICIAL!
We're announcing our partnership with — a privacy-first, uncensored AI inference platform on — as a primary inference API backend for StrikeRobot’s product offerings.
The partnership involves Venice's credit sponsorship and co-development with
2:00 AM · May 7, 2026 · 86.1K Views
57 Replies · 76 Reposts · 437 Likes
Hermes Agent v0.13.0 launched with multi-agent orchestration via a Kanban system, enforced goal completion, and custom LLM provider support
Introducing Hermes Agent v 0.13.0
- Multi-Agent orchestration through the Kanban system
- Enforced goal completion with /goal
- Big optimizations for disk usage
- Much more extensibility, custom LLM Providers, custom gateway channels, and much more
Nous Research @NousResearch
Hermes Agent v0.13.0 - “The Tenacity Release”
9:05 PM · May 7, 2026 · 536K Views
68 Replies · 90 Reposts · 1.12K Likes
Crypto Integrated is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
