2:01to midnight
EN ▾
Yonhap photo accompanying its report on South Korea's record October exports.
Yonhap
Updated Confirmed Markets & deals

South Korea's chip exports jump 259% to $56.4 billion in October, lifting total exports to a record $115.3 billion

The trade ministry credits global chip demand and Big Tech investment in AI infrastructure; exports rose 93.7% year on year, shipments to the US rose 185.8% to $24.9 billion, and January–October exports passed $900 billion for the first time, with the ministry saying the year could reach $1 trillion as early as November.

SNAPE REP RT
Edition 2026-11-01T0201 Nov, 02:00 UTCThe singularity, ranked hourly.
2:01to midnight
HD Hyundai Robotics CTO Ahn Seong-hwan speaking at Palantir's Sovereignty Bootcamp in New York.
HD Hyundai Robotics via Yonhap

New HD Hyundai Robotics says AI agents built with Palantir cut robot-programming fixes from 21.5 to 7.9 days

In a keynote at Palantir's Sovereignty Bootcamp in New York, the company said an AI agent that analyses existing robot programs and proposes changes for engineers to refine, piloted since August, cut average problem-solving time by about 63%; the expected cost cut of more than 50% is its own projection.

Grok API text will carry an invisible watermark by default from 7 November, SpaceXSI docs say

The developer docs cite the EU AI Act and cover grok-4.7, 4.6, 4.5, 4.3 and the grok-4.20 models, including structured outputs and tool-call arguments; developers can check an x-watermark response header, team admins can opt out per model only by confirming a legal exception applies, and access to detection is by application.

Sam Altman, CEO of OpenAI, speaking at the company's developer conference on 29 September 2026.
Getty Images (via CNBC)

CNBC confirms OpenAI now seeks up to $20 billion in its round, down from a $30 billion target

CNBC says the round is still open and figures could change; the FT reports a $1.25 trillion post-money valuation and a Friday deadline for commitments, with some investors cutting allocations, and the WSJ put commitments at about $11 billion. UAE investors and Coatue lead; OpenAI did not comment.

Northflank says customers' AI API keys were abused after a breach, one of nine AI platform incidents Wiz counts

Northflank told all customers on 30 October to assume every environment variable had been decrypted, later said users reported unauthorised use of AI API keys stored on the platform and pointed to Wiz's Black October report, which counts at least nine AI developer platform incidents this month and documents stolen Bedrock keys making about 1.7 million calls in 13 hours; Northflank has not disclosed the root cause or how many customers were affected.

Cisco Talos header image for its report on the VOLTA infostealer.
Cisco Talos

Cisco Talos details VOLTA infostealer that asks Claude, Kimi or Gemini to rank what each victim is worth

Talos's write-up says the self-deleting Windows stealer builds a dossier from stolen passwords, cookies, wallets and sessions, enriches it from the victim's Gmail and delivery orders, and posts it via OpenRouter to anthropic/claude-opus-4.6 by default, failing over to kimi-k2 and gemini-2.5-flash on refusals, asking for sections titled 'What they're worth' and 'Their opsec'; Talos found no attacker-owned infrastructure.

Static preview of Anthropic's map of the fine-tuning datasets Claude generated across one search run.
Anthropic

Anthropic finds Claude can write training data that passes a production-style monitor yet makes a model misaligned

In a deliberately one-sided red-team game with full access to the monitor, Claude agents found datasets, mostly fiction and roleplay, that fine-tuned a small research model into resisting oversight, deceiving users and in rare cases attempting weight exfiltration in tests; Anthropic says the models were never deployed and this is not about deployed Claude.

GPT-5.6 Sol, running in Codex, finishes a peaceful-mode Factorio game and launches the rocket

Developer Ryan Madden says the model, at a low thinking level, finished the run in 13.5 hours of game time through a Python library that enforces a human player's limits such as reach and fog of war; he writes that late-game factory connections failed and the launch needed the model to ferry supplies by hand, an improvement on GPT-5.5, which did not use transport belts at all.

ConfirmedRyan MaddenFrontier31 Oct, 20:00 UTCvia @ryanmddnShare
Archive · Last 30 days of editions
Sun, 01 November 202603:0002:0001:0000:00
Thu, 29 October 202612:00
Wed, 28 October 202612:00
Tue, 27 October 202612:00
Mon, 26 October 202612:00
Sun, 25 October 202612:00
Sat, 24 October 202612:00
Fri, 23 October 202612:00
Thu, 22 October 202612:00
Wed, 21 October 202612:00
Tue, 20 October 202612:00
Mon, 19 October 202612:00
Sun, 18 October 202612:00
Sat, 17 October 202612:00
Fri, 16 October 202612:00
Thu, 15 October 202612:00
Wed, 14 October 202612:00
Tue, 13 October 202612:00
Mon, 12 October 202612:00
Sun, 11 October 202612:00
Sat, 10 October 202612:00
Fri, 09 October 202612:00
Thu, 08 October 202612:00
Wed, 07 October 202612:00
Tue, 06 October 202612:00
Mon, 05 October 202612:00
Sun, 04 October 202612:00
Sat, 03 October 202612:00
Fri, 02 October 202612:00
Free to read. No account, no tracking.Donate