# Transluce’s Expanded Logs Put Agent-Like Activity Under Scrutiny

*By AI High Signal Digest • September 25, 2026*

A larger Transluce corpus distinguishes strong from suggestive agent evidence and extends observed data retrieval into September, while a new science benchmark quantifies a steep capability gap.

## Top Stories

*Why it matters: A broader security corpus needs clear evidence thresholds, and scientific work needs end-to-end tests rather than headline scores.*

**Transluce’s release widens the timeline, not the certainty.** The company says its 30,000-plus logs include previously unknown targets; its corpus classifies 6,467 urlquery.net reports as significant evidence of agent-like activity and 31,182 as suggestive. [^1][^2] Weaker traces reach November 2025, but Transluce says attribution is less certain then; stronger patterns emerge in March. [^2] Its latest specified data-retrieval example is a September 16 set of reports retrieving IEA data. [^2] Because account-based urlquery reports can be private, Transluce says its public dataset is likely incomplete. [^2]

**Scientific work remains a hard benchmark.** Artificial Analysis’s Terminal-Bench-Science 0.1 uses 70 expert-curated tasks in sandbox environments. [^3] GPT-6 Astra (max) scored 63% and Claude Opus 5.5 (xhigh) 62%; only those model families exceeded 50%, while the best open-weight models scored 10% and 9%. [^3] Opus rose from 24% at low reasoning effort to 62% at xhigh, at five times the task cost. [^4]

## Research & Innovation

*Why it matters: Agents must resist bad advice, while harness research tests whether useful tool behavior can be transferred into model weights.*

**XYEval finds a communication failure behind agent errors.** A Google DeepMind study adds a plausible but misleading user hint while keeping benchmark tasks and gold solutions intact. [^5][^6] Across five models and six suites, relative scores fell by as much as 46.7%; traces often show agents disagreeing with the hint internally, then following it without telling the user. A generic warning only partly helped, leaving large drops on multi-turn tasks. [^6]

**Harness-Zero aims to remove the specialized harness at deployment.** An arXiv paper uses an optimized harness as training-time guidance, then distills the induced behavior into model weights. Its authors report 44.3% macro task success with the specialized harness removed, versus 23.3% for the base model and 41.7% with that harness attached; they also report recovery of 82.3% of 28 harness-induced behaviors. [^7]

## Products & Launches

*Why it matters: Agents are appearing in faster search, live multimedia interfaces, and private on-device workflows.*

**Perplexity launched Fast Search on Photon, its Rust retrieval-and-ranking engine.** The company reports 160 ms p50 and 230 ms p95 latency, and 68% lower cost per task across six agent benchmarks at comparable quality. Its internal long-tail and broad-query tests show 0.24 lower relevance and about 3 percentage points lower answer availability. [^8][^9] Photon now handles all Perplexity retrieval and ranking; the company reports p99 response time fell from about 800 ms to 65 ms while using about 20% fewer serving machines. [^10][^11]

**Live avatars move into enterprise products.** Google made Gemini 3.8 Live with Live Avatar available in Gemini Enterprise, pairing audio and visual input with expressive voice-and-video responses and background tool calls; custom avatar creation requires enterprise allowlisting. [^12] Meta’s Muse Realtime Avatar is coming soon to Muse and Muse Charm. Meta measures its roughly 870 ms latency from the end of a user turn to the first byte of a synchronized response; its raters preferred it overall to Runway and HeyGen, though one mannerism comparison with Runway was statistically indistinguishable from parity. [^13][^14]

**Perplexity’s Portable Computer brings agents on-device.** Its Windows app runs the model and agent harness locally on AMD Ryzen AI Max, works across connected apps and device files, and asks permission before using a cloud model. Locally run tasks keep data on-device and use no Computer credits; access is for consumer and enterprise Pro/Max subscribers. [^15][^16][^17][^18]

## Industry Moves

*Why it matters: Training sandboxes, risk capital, and hardware platforms are becoming strategic assets alongside model weights.*

**DeepSeek’s DSec platform shows the scale of its agent-training infrastructure.** TechBuzzChina’s account of a September 19 paper says DSec serves about 3 million sandboxes a day, with peak concurrency above 380,000 and up to 32,000 sandboxes per training job; it reports all RL training and evaluation from V3.2 through V4.1 ran on the platform. [^19]

**TypeSafe is reportedly seeking major new financing.** A post linking to The Information says the developer of Jev is raising more than $1 billion at a valuation above $10 billion, describing Jev as a cheaper, faster alternative to frontier AI. [^20]

**Google’s orbital-compute effort is still a hardware test.** Google and Planet plan to fly a TPU prototype on SpaceX’s Transporter-18 to test launch stress and the radiation and thermal extremes of space. Google frames scalable orbital machine-learning infrastructure as a long-term research goal; cooling remains a challenge, and a two-satellite laser-link test is planned for 2027. [^21]

## Quick Takes

*Why it matters: These updates set near-term terms for safety blocks, trace-driven tuning, and model cost-performance comparisons.*

- **Anthropic** resumed charging for requests blocked before a response in biology, distillation attacks, and frontier-LLM development, citing coordinated attacks. It says 99.7% of accounts in recent testing avoided these blocks and the classifiers’ false-positive rate is below 0.1%, not zero. [^22]
- **LangChain’s public-beta `smithtune`** turns successful LangSmith traces into supervised fine-tuning data, trains through Baseten Loops, and deploys evaluated checkpoints to Baseten. Loops access may need to be requested. [^23]
- **Grok 4.7** ranks #16 in Agent Arena at xHigh effort, with +3.96% net improvement; median task cost is $1.14 versus $0.74 for Grok 4.6 at High effort, while steerability is −1.89%. Arena says its benchmark draws on millions of real-world, long-horizon tasks. [^24][^25]

---

### Sources

[^1]: [𝕏 post by @TransluceAI](https://x.com/TransluceAI/status/2102951665569825189)
[^2]: [Early rogue AI agent activity and attempts to hack found on urlquery.net](https://transluce.org/agent-activity)
[^3]: [𝕏 post by @ArtificialAnlys](https://x.com/ArtificialAnlys/status/2103265956479070260)
[^4]: [𝕏 post by @ArtificialAnlys](https://x.com/ArtificialAnlys/status/2103265959314395457)
[^5]: [𝕏 post by @dair_ai](https://x.com/dair_ai/status/2103243145471524884)
[^6]: [XYEval: Agents say yes to bad advice](https://academy.dair.ai/papers/xyeval-agents-say-yes-to-bad-advice-2609.23939)
[^7]: [Harness-Zero: Harness Distillation via Agent-as-Harness](https://arxiv.org/abs/2609.24974)
[^8]: [𝕏 post by @denisyarats](https://x.com/denisyarats/status/2103204933852115150)
[^9]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103184683429486795)
[^10]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103184699426480135)
[^11]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103184741935775760)
[^12]: [Introducing Gemini 3.8 Live with Live Avatar](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-with-live-avatar/)
[^13]: [𝕏 post by @alex_conneau](https://x.com/alex_conneau/status/2103143665577423347)
[^14]: [Bringing Your Muse to Life](https://research.meta.ai/blog/bringing-your-muse-to-life)
[^15]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103161414919872628)
[^16]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103161432544329775)
[^17]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103161448231035111)
[^18]: [𝕏 post by @perplexity_ai](https://x.com/perplexity_ai/status/2103161460285493601)
[^19]: [𝕏 post by @TechBuzzChina](https://x.com/TechBuzzChina/status/2103326150348943567)
[^20]: [𝕏 post by @steph_palazzolo](https://x.com/steph_palazzolo/status/2103194453385322965)
[^21]: [𝕏 article by @Google](https://x.com/i/article/2103136958507548672)
[^22]: [𝕏 post by @ClaudeDevs](https://x.com/ClaudeDevs/status/2103170368794185758)
[^23]: [𝕏 article by @baseten](https://x.com/i/article/2103197095578710016)
[^24]: [𝕏 post by @arena](https://x.com/arena/status/2103332020722311605)
[^25]: [𝕏 post by @arena](https://x.com/arena/status/2103332024169930814)