Techmeme

#Techmeme

Techmeme 20260430 Anthropic Plan to Expand Mythos Access Is Opposed by White House Summary

2026-04-30 911 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 30, 2026 story in its Anthropic Mythos cluster, and the direct source used here is Bloomberg’s Anthropic Plan to Expand Mythos Access Is Opposed by White House.

The White House is opposing Anthropic’s plan to broaden access to Claude Mythos Preview, the company’s restricted frontier model for advanced cybersecurity work. Anthropic had proposed granting Mythos access to roughly 70 more companies and organizations, expanding a program that already includes launch partners and dozens of critical software maintainers under Project Glasswing. Administration officials objected on security grounds, and the Techmeme cluster also highlighted Bloomberg’s separate report that the NSA has been testing Mythos to find vulnerabilities in Microsoft products and other widely used software.

Continue ...

Techmeme 20260424 DeepSeek-V4 Towards Highly Efficient Million-Token Context Intelligence Summary

2026-04-24 883 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 24, 2026 story in its Techmeme item, and the direct source used here is DeepSeek’s DeepSeek-V4 technical report.

DeepSeek released preview versions of DeepSeek-V4-Pro and DeepSeek-V4-Flash, two Mixture-of-Experts models aimed at a specific claim: open models can push much farther into long-context and agentic work without the usual explosion in cost. Pro is a 1.6T-parameter model with 49B activated parameters, while Flash is 284B total with 13B activated. Both support one-million-token contexts and were trained on more than 32T tokens.

Continue ...

Techmeme 20260424 Project Deal Our Claude-Run Marketplace Experiment Summary

2026-04-24 869 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced Anthropic’s April 24, 2026 post, Project Deal: our Claude-run marketplace experiment, as a concrete example of agent-to-agent commerce moving out of theory and into something closer to a real market.

Anthropic set up a one-week internal classified marketplace for 69 employees in its San Francisco office and let Claude agents negotiate on both sides of each transaction. Employees told Claude what they might want to buy or sell, gave it some constraints and style guidance, and then stepped out of the loop. Each participant got a $100 budget, the agents ran across parallel Slack channels, and any deal reached by the agents was later honored by the humans in person.

Continue ...

Techmeme 20260423 Bad Connection Uncovering Global Telecom Exploitation by Covert Surveillance Actors Summary

2026-04-23 704 words 4 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 25, 2026 story through TechCrunch’s report, and the direct source used here is Citizen Lab’s Bad Connection: Uncovering Global Telecom Exploitation by Covert Surveillance Actors.

Citizen Lab says it uncovered two sophisticated telecom-surveillance campaigns that exploited the global signalling systems connecting mobile carriers, combining older SS7 pathways with Diameter traffic used in 4G and most 5G roaming. One campaign targeted a “VVIP” company executive across multiple 3G and 4G networks. Another used a specially formatted SMS carrying hidden SIM card commands intended to turn the victim’s device into a tracking beacon. The larger point is that both campaigns treated the mobile network itself as the surveillance platform.

Continue ...

Techmeme 20260423 Introducing GPT-5.5 Summary

2026-04-23 957 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 23, 2026 story in its Techmeme item, and the original article is OpenAI’s Introducing GPT-5.5.

OpenAI is positioning GPT-5.5 as more than a routine model refresh. The company says the model is better at carrying work forward with less micromanagement: understanding messy instructions, planning steps, using tools, moving between apps, checking results, and finishing longer tasks instead of stalling halfway through them. In practical terms, OpenAI is aiming GPT-5.5 at coding, browser-based research, data analysis, documents, spreadsheets, computer use, and early-stage scientific work.

Continue ...

Techmeme 20260422 Making ChatGPT Better for Clinicians Summary

2026-04-22 1009 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 22, 2026 story in its Techmeme item, and the original article is OpenAI’s Making ChatGPT better for clinicians.

OpenAI is moving ChatGPT more directly into clinical work with a new product called ChatGPT for Clinicians. The product is designed for tasks like documentation, medical research, care consults, referral letters, prior authorization support, and patient instructions. It is initially free for verified individual clinicians in the United States, including physicians, nurse practitioners, physician assistants, and pharmacists.

Continue ...

Techmeme 20260421 SpaceX Is Working with Cursor and Has an Option to Buy the Startup for 60B Summary

2026-04-21 990 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this story on its April 22, 2026 page, and the original article is TechCrunch’s SpaceX is working with Cursor and has an option to buy the startup for $60B.

At the center of the story is a deal that is much bigger than a normal infrastructure partnership. TechCrunch reports that SpaceX told investors it is working with Cursor on a next-generation coding and knowledge-work AI effort, and that the agreement gives SpaceX a choice later this year: either pay Cursor $10 billion for the partnership or acquire the company for $60 billion. Even in a market that has become numb to giant AI numbers, that structure is startling. It treats the collaboration itself as enormously valuable while also creating a path to full vertical integration if the training work goes well enough.

Continue ...

Techmeme 20260420 Kimi K2.6 Tech Blog Advancing Open-Source Coding Summary

2026-04-20 1004 words 5 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 20, 2026 story, and the original post is Kimi K2.6 Tech Blog: Advancing Open-Source Coding.

Moonshot is positioning Kimi K2.6 as more than a routine model refresh. The company says it is open sourcing a new model centered on long-horizon coding, agent-style execution, coding-driven design, proactive agents, and a research preview called Claw Groups. In practical terms, the launch is framed as an attempt to make open models feel capable not just in isolated code generation tasks, but across the broader workflow that increasingly defines modern AI-assisted software work: planning, tool use, extended sessions, parallel sub-agents, and end-to-end execution.

Continue ...

Techmeme 20260419 Cloud Development Platform Vercel Was Hacked Summary

2026-04-19 813 words 4 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 19, 2026 story, and the direct article used here is The Verge’s Cloud development platform Vercel was hacked, which also points to Vercel’s official security bulletin.

Vercel says it identified unauthorized access to certain internal systems and has brought in incident-response specialists while notifying law enforcement. The company says its services remain operational and that, so far, it has identified only a limited subset of customers as affected. Those customers are being contacted directly.

Continue ...

Techmeme 20260416 Salesforce Launches Headless 360 to Turn Its Entire Platform into Infrastructure for AI Agents Summary

2026-04-18 769 words 4 minutes #techmeme

Generated by Codex with GPT-5

What happened

Techmeme surfaced this April 16, 2026 story, and the original article is VentureBeat’s Salesforce launches Headless 360 to turn its entire platform into infrastructure for AI agents.

Salesforce is making a blunt strategic argument: in an agentic software world, the browser should stop being the main doorway into enterprise software. Headless 360 is its attempt to turn Salesforce from an application that humans click through into a programmable substrate that agents can operate directly. Instead of forcing every workflow through a web UI, the company now exposes platform capabilities as APIs, MCP tools, and CLI commands.

Continue ...