# get merge: Friday, July 3, 2026

17 stories merged from 4 newsletters.

AI & design news, merged daily. One briefing that merges and deduplicates stories from leading AI and design newsletters.

Web: https://getmerge.ai/2026-07-03

## TLDR

- OpenAI floats 5% government stake as Altman pitches global safety forum; Fable 5 returns after 19-day security timeout; Claude Science beta launches with 60+ research databases.
- LLMs have broken legibility of effort; NVIDIA splits 30B model for 2.42x faster generation; Hugging Face + Cerebras ship open real-time voice AI.
- Nous Research Hermes Agent v0.18.0; GitHub spec-kit; Seedance 2.5 and Cursor for iOS among seven new tools.

## Top Stories

- **OpenAI floats 5% government stake; Altman pitches global safety forum**. OpenAI reportedly discussed giving the U.S. government a 5% equity stake for public wealth-sharing, while Sam Altman called for a U.S.-led international AI safety forum modeled on the IAEA. Sources: The Neuron, The Rundown AI. [1](https://www.cnbc.com/2026/07/02/openai-proposes-us-government-own-5percent-stake-to-address-political-blowback.html) [2](https://www.ft.com/content/0c2e1077-f658-4b3d-9040-602615c961ca)
- **Anthropic Claude Fable 5 returns after 19-day security timeout**. Amazon researchers found a bypass that triggered a national-security pull of Fable 5. Anthropic added a classifier blocking 99%+ of the reported technique, routes some requests to Opus 4.8, and is drafting a cross-industry jailbreak framework. Sources: AlphaSignal, The Neuron. [3](https://www.anthropic.com/news/redeploying-fable-5)
- **Claude Science beta launches with 60+ research databases**. Anthropic's dedicated research app connects natively to UniProt, PDB, ChEMBL and more. Includes a reviewer agent for citation/figure verification. UCSF researchers report ~10x speedup on certain analyses. Sources: AlphaSignal. [4](https://claude.com/product/claude-science)
- **Microsoft launches Frontier Company: $2.5B enterprise AI unit**. Microsoft committed $2.5B and 6,000 engineers to move enterprise AI from pilots to production systems, placing specialists at client sites. Sources: The Neuron, The Rundown AI. [5](https://blogs.microsoft.com/blog/2026/07/02/microsoft-frontier-company-ai-engineering-that-amplifies-and-protects-your-intelligence/)
- **TML + Bridgewater: specialized 235B model beats frontier giants on financial tasks**. Thinking Machines Lab and Bridgewater trained Qwen3-235B on expert-graded examples, achieving 84.7% accuracy on financial tasks vs ~50% for GPT/Claude/Gemini at 13.8x lower cost. Sources: The Rundown AI. [6](https://thinkingmachines.ai/news/learning-to-replicate-expert-judgment-in-financial-tasks/)

## Deeper Cuts

- **LLMs have broken legibility of effort**. How generative AI has destroyed our ability to tell whether something took real human work, impacting craft, trust, and creative labor valuation. Sources: Sidebar. [7](https://eieio.games/blog/legibility-of-effort/)
- **NVIDIA splits 30B model for 2.42x parallel generation**. Nemotron-Labs-TwoTower writes text in chunks rather than one token at a time. 98.7% quality retained on benchmarks. Runs on 2x H100 or A100. Sources: AlphaSignal. [8](https://huggingface.co/nvidia/Nemotron-Labs-TwoTower-30B-A3B-Base-BF16)
- **Hugging Face + Cerebras ship open real-time voice AI**. A fully open speech-to-speech assistant with replaceable modules for listening, thinking, and speaking. Free demo and repo available. Sources: The Neuron. [9](https://huggingface.co/blog/cerebras-gemma4-voice-ai)
- **White House model-release standards talks accelerate**. Frontier labs and national-security agencies negotiate voluntary release benchmarks and access rules, gaining urgency after June's Mythos/Fable 5 incidents. Sources: The Neuron. [2](https://www.ft.com/content/0c2e1077-f658-4b3d-9040-602615c961ca)
- **Anthropic in early custom-chip talks with Samsung**. After poaching Clive Chan (lead behind OpenAI's Jalapeño chip), Anthropic reportedly approached Samsung about manufacturing its own AI chip. Sources: The Rundown AI. [10](https://www.theinformation.com/articles/anthropic-talks-samsung-manufacture-custom-ai-chip)

## Tools and Launches

- **Nous Research Hermes Agent v0.18.0**. Open-source agent framework with new judgment capabilities. Sources: AlphaSignal. [11](https://github.com/NousResearch/hermes-agent)
- **GitHub spec-kit**. Open-source tool that structures AI coding work via a spec-first workflow before code is written. Sources: AlphaSignal. [12](https://github.com/github/spec-kit)
- **Claude Tag for Slack**. Tag @Claude as a teammate inside Slack threads for cloud-based task delegation with memory and tool access. Sources: The Rundown AI. [13](https://www.anthropic.com/news/introducing-claude-tag)
- **Safari MCP Server**. Lets AI agents connect to a real Safari Tech Preview browser for inspection, screenshots, logs, and debugging. Sources: The Neuron. [14](https://webkit.org/blog/18136/introducing-the-safari-mcp-server-for-web-developers/)
- **LangBot**. Open-source platform to deploy AI bots across Slack, Discord, and WeChat. Sources: AlphaSignal. [15](https://github.com/langbot-app/LangBot)
- **Seedance 2.5**. ByteDance upgraded video model for 30-second cinematic clips with R2V control. Sources: The Neuron, The Rundown AI. [16](https://dreamina.capcut.com/seedance/seedance-2-5)
- **Cursor for iOS**. Mobile coding experience from the AI-native editor. Sources: The Rundown AI. [17](https://www.rundown.ai/tools/cursor-for-ios)

## References

1. [CNBC: OpenAI proposes US government 5% stake](https://www.cnbc.com/2026/07/02/openai-proposes-us-government-own-5percent-stake-to-address-political-blowback.html)
2. [FT: Altman pitches US-led AI safety forum](https://www.ft.com/content/0c2e1077-f658-4b3d-9040-602615c961ca)
3. [Anthropic: Redeploying Fable 5](https://www.anthropic.com/news/redeploying-fable-5)
4. [Claude: Claude Science product page](https://claude.com/product/claude-science)
5. [Microsoft: Frontier Company announcement](https://blogs.microsoft.com/blog/2026/07/02/microsoft-frontier-company-ai-engineering-that-amplifies-and-protects-your-intelligence/)
6. [Thinking Machines Lab: TML + Bridgewater research](https://thinkingmachines.ai/news/learning-to-replicate-expert-judgment-in-financial-tasks/)
7. [eieio.games: Legibility of Effort](https://eieio.games/blog/legibility-of-effort/)
8. [Hugging Face: NVIDIA Nemotron-Labs-TwoTower-30B](https://huggingface.co/nvidia/Nemotron-Labs-TwoTower-30B-A3B-Base-BF16)
9. [Hugging Face: Cerebras voice AI](https://huggingface.co/blog/cerebras-gemma4-voice-ai)
10. [The Information: Anthropic Samsung chip talks](https://www.theinformation.com/articles/anthropic-talks-samsung-manufacture-custom-ai-chip)
11. [Nous Research: Hermes Agent](https://github.com/NousResearch/hermes-agent)
12. [GitHub: spec-kit](https://github.com/github/spec-kit)
13. [Anthropic: Introducing Claude Tag](https://www.anthropic.com/news/introducing-claude-tag)
14. [WebKit: Safari MCP Server](https://webkit.org/blog/18136/introducing-the-safari-mcp-server-for-web-developers/)
15. [GitHub: LangBot](https://github.com/langbot-app/LangBot)
16. [Dreamina: Seedance 2.5](https://dreamina.capcut.com/seedance/seedance-2-5)
17. [The Rundown: Cursor for iOS](https://www.rundown.ai/tools/cursor-for-ios)

Sources today: [AlphaSignal](https://alphasignal.ai/), [The Neuron](https://www.theneurondaily.com/), [The Rundown AI](https://www.therundown.ai/), [Sidebar](https://sidebar.io/).