2026-05-08 — Linnet Daily

§ 0

The Morning

Local weather 1

This morning in

London

Partly cloudy

Today's range

18.1°↓9.2°

currently 15.9°

Feels

13.1°

Rain

5%

Wind

9 km/h

Humid

44%

Rise

05:19

Set

20:34

§ I

US Stocks

Pre-market signal radar 12

US pre-market radar

premarket 2026-05-08

10 Bullish

0 Bearish

2 Neutral

Sector Tape

Compute Mining 4 names

78 Top: IREN · Bullish · RS +1.9% Bullish 3 / Bearish 0 / 5d +21.4%

Foundry 2 names

74 Top: INTC · Bullish · RS +3.6% Bullish 2 / Bearish 0 / 5d +10.3%

Servers and Thermal Management 2 names

73 Top: VRT · Bullish · RS +0.4% Bullish 2 / Bearish 0 / 5d +6.8%

Networking Equipment 4 names

68 Top: CRDO · Neutral · RS -10.1% Bullish 1 / Bearish 0 / 5d -3.7%

Hyperscale Cloud 4 names

68 Top: ORCL · Neutral · RS +1.1% Bullish 0 / Bearish 0 / 5d +7.4%

Manufacturing 4 names

66 Top: FLEX · Neutral · RS +6.8% Bullish 1 / Bearish 0 / 5d +9.9%

Battery and Energy Storage 3 names

56 Top: FLNC · Neutral · RS +13.8% Bullish 1 / Bearish 0 / 5d +11.8%

Energy Infrastructure 1 names

48 Top: VST · Neutral · RS +2.5% Bullish 0 / Bearish 0 / 5d -2.5%

Ticker Setup Move Score Evidence Quality

IREN IREN Ltd Compute Mining

Bullish Gap up + news High confidence

+8.0% $61.40 5d +24.9%

85 sector positive RS +5.4%

Bullish setup from +8.0% vs previous close, positive sector tape, 3 recent headline(s).

Spain deal gives IREN 490MW as AI cloud push moves into Europe - Stock Titan Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

WULF TeraWulf Compute Mining

Bullish Gap up + news High confidence

+4.1% $25.00 5d +10.5%

81 sector positive RS -9.0%

Bullish setup from +4.1% vs previous close, positive sector tape, 3 recent headline(s).

TeraWulf (WULF) Valuation Check After AI Shift And US$1.0b Stock Offering - simplywall.st Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

INTC Intel Foundry

Bullish Gap up + news High confidence

+1.3% $111.07 5d +16.0%

78 sector positive RS +9.4%

Bullish setup from +1.3% vs previous close, positive sector tape, 3 recent headline(s).

Intel Exec Sells $4 Million in Stock After Record-Breaking Rally - Barron's Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

FLEX Flex Ltd Manufacturing

Bullish Sector tailwind High confidence

+0.9% $134.24 5d +45.3%

77 sector positive RS +42.2%

Bullish setup from +0.9% vs previous close, positive sector tape, 3 recent headline(s).

QRG Capital Management Inc. Boosts Stock Holdings in Flex Ltd. $FLEX - MarketBeat Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

CRDO Credo Technology Networking Equipment

Bullish Gap up + news High confidence

+1.6% $191.22 5d +8.2%

76 sector positive RS +1.8%

Bullish setup from +1.6% vs previous close, positive sector tape, 3 recent headline(s).

Geneva Partners LLC Takes $4.36 Million Position in Credo Technology Group Holding Ltd. $CRDO - MarketBeat Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

HUT Hut 8 Compute Mining

Bullish Gap up + news Medium confidence

+1.2% $102.36 5d +33.5%

75 sector positive RS +14.0%

Bullish setup from +1.2% vs previous close, positive sector tape, 3 recent headline(s).

Hut 8 (NASDAQ:HUT) Shares Down 11.8% - Here's Why - MarketBeat Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

VRT Vertiv Holdings Servers and Thermal Management

Bullish Gap up + news High confidence

+2.0% $346.71 5d +3.5%

75 sector positive RS -3.0%

Bullish setup from +2.0% vs previous close, positive sector tape, 3 recent headline(s).

Vertiv Hires New CPO To Tackle AI Data Center Supply Chain - simplywall.st Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

DELL Dell Technologies Servers and Thermal Management

Bullish Gap up + news Medium confidence

+1.4% $233.50 5d +10.2%

71 sector positive RS +3.7%

Bullish setup from +1.4% vs previous close, positive sector tape, 3 recent headline(s).

DELL Stock Price, Quote, Chart & Forecast - Techi Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

FLNC Fluence Energy Battery and Energy Storage

Bullish Gap up + news Medium confidence

+25.9% $23.88 5d +55.8%

71 sector negative RS +57.7%

Bullish setup from +25.9% vs previous close, negative sector tape, 3 recent headline(s).

Fluence Energy Stock Surges 60%, With A 6-Day Winning Spree - Trefis Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

TSM Taiwan Semiconductor Foundry

Bullish Sector tailwind Medium confidence

+0.1% $414.60 5d +4.6%

70 sector positive RS -2.1%

Bullish setup from positive sector tape, 3 recent headline(s).

Taiwan Semiconductor Manufacturing Company Ltd. $TSM Stock Position Reduced by GRIMES & Co WEALTH MANAGEMENT LLC - MarketBeat Weakens if price fades below previous close or sector benchmarks roll over.

quote: delayed fallback news: fresh financials: fresh news: 3

ORCL Oracle Hyperscale Cloud

Neutral Sector tailwind Medium confidence

-1.6% $191.42 5d +20.6%

70 sector positive RS +14.3%

Watchlist item from -1.6% vs previous close, positive sector tape, 3 recent headline(s).

Oracle Stock 5-Day Winning Spree: Stock Climbs 21% - Trefis Needs fresh price/news confirmation before becoming an actionable setup.

quote: delayed fallback news: fresh financials: fresh news: 3

APH Amphenol Networking Equipment

Neutral Gap up + news Medium confidence

+1.0% $138.00 5d -7.2%

69 sector positive RS -13.6%

Watchlist item from +1.0% vs previous close, positive sector tape, 3 recent headline(s).

Generali Asset Management SPA SGR Acquires 14,612 Shares of Amphenol Corporation $APH - MarketBeat Needs fresh price/news confirmation before becoming an actionable setup.

quote: delayed fallback news: fresh financials: fresh news: 3

quotes: nasdaq 24 24/24news: google_news_rss 24 24/24filings: sec 24 24/24, fallback 24

Generated from public market data and news for research and education. Not financial advice; data may be delayed, incomplete, or wrong.

§ II

From the arXiv

arXiv preprints 10 of 20

cs.AIarxiv:2605.06638v1Lead article

Can RL Teach Long-Horizon Reasoning to LLMs? Expressiveness Is Key

Tianle Wang, Zhaoyang Wang, Guangchen Lan, Xinpeng Wei, Sipeng Zhang

his paper introduces **ScaleLogic**, a synthetic framework to systematically study how Reinforcement Learning (RL) improves LLM reasoning across varying proof depths (horizon) and logical expressiveness. The core contribution is demonstrating that the required RL training compute scales with reasoning depth via a power law, where the scaling exponent increases significantly as the underlying logic becomes more expressive (e.g., incorporating "and," "or," and "not").

Read abstract →Full PDF

Overview of ScaleLogic . Each problem has B B candidate proof trees, exactly one of which has a provable conclusion; the others are made unprovable by corrupting one axiom. The depth D D controls proof depth. Left: Implication-only reasoning. Right: The most expressive logic setting (referred to as + Quantification in Section 3.2 ) combines conjunction, disjunction, negation, and universal quantification. — Overview of ScaleLogic . Each problem has B B candidate proof trees, exactly one of which has a provable conclusion; the others are made unprovable by corrupting one axiom. The depth D D controls proof depth. Left: Implication-only reasoning. Right: The most expressive logic sett…

The Overall Workflow of Cola DLM. Detailed illustration of the training and inference pipeline of 𝒞 o l a $ \mathcal{C} $ola DLM . Training Stage 1 shows Text VAE pretraining with reconstruction, BERT, and KL losses. Training Stage 2 shows joint pretraining of the Text VAE and Text DiT with gradient control for stable optimization, where a specialized block-causal mechanism is adopted in the DiT. Inference Stage illustrates the decoding process with KV cache. — The Overall Workflow of Cola DLM. Detailed illustration of the training and inference pipeline of 𝒞 o l a $ \mathcal{C} $ola DLM . Training Stage 1 shows Text VAE pretraining with reconstruct…

cs.AIarxiv:2605.06548v1

Continuous Latent Diffusion Language Model

Hongcan Guo, Qinyu Zhao et al.

This paper introduces Cola DLM, a hierarchical latent diffusion language model that decomposes text generation into distinct stages. It first maps text to a stable latent space using a Text VAE, then models a global semantic prior using a block-causal DiT in t…

abstract pdf

cs.AIarxiv:2605.06490v1

Instrumental Choices: Measuring the Propensity of LLM Agents to Pursue Instrumental Behaviors

Jonas Wiedermann-Möller, Leonard Dung et al.

This paper introduces "Instrumental Choices," a benchmark to measure the propensity of LLM agents to engage in instrumental convergence (IC) behaviors, such as self-preservation, which might lead to instruction violation for goal utility. The benchmark uses se…

abstract pdf

Aggregate adjusted instrumental-convergence (IC) behaviour rate by model over all tasks and variants ( n = 168 n=168 samples per model). Error bars show 95% Wilson confidence intervals over sample-level adjusted IC labels. — Aggregate adjusted instrumental-convergence (IC) behaviour rate by model over all tasks and variants ( n = 168 n=168 samples per model). Error bars show 95% Wilson confidence intervals over sample-lev…

Overview of the MASPO Framework. The optimization proceeds sequentially following the topological order of the agent graph (Top-Right). (Top) For a specific target agent, the Prompt Optimizer analyzes execution traces (context 𝒞 $ \mathcal{C} $ and output o o ) from sampled batches ℬ i t e r ∪ ℬ m i s $ \mathcal{B}_{iter} $$ \cup $$ \mathcal{B}_{mis} $ to generate candidate prompts 𝒫 c a n d $ \mathcal{P}_{cand} $ . These candidates are rigorously assessed by the LLM Evaluator across three distinct dimensions: local adherence, lookahead potential, and global alignment. (Bottom-Left) To resolve credit assignment, we synthesize these evaluations into a Joint Reward Model. Crucially, we identify and mine Misalignment Cases to explicitly guide the optimizer towards repairing coordination breakdowns. (Bottom-Right) Navigating the high-dimensional search space, the framework employs a Trace-Guided Beam Search. This mechanism maintains a beam of Top-K candidates, accumulating joint reward scores along the path to iteratively evolve and select the optimal prompt. — Overview of the MASPO Framework. The optimization proceeds sequentially following the topological order of the agent graph (Top-Right). (Top) For a specific target agent, the Prompt Optimizer analyzes…

cs.AIarxiv:2605.06623v1

MASPO: Joint Prompt Optimization for LLM-based Multi-Agent Systems

Zhexuan Wang, Xuebo Liu et al.

MASPO is a novel framework for jointly optimizing role-specific prompts in LLM-based Multi-Agent Systems. Its core method involves a joint evaluation mechanism that assesses prompts based on their contribution to downstream agent success, bridging local and gl…

abstract pdf

cs.AIarxiv:2605.06584v1

NeuroAgent: LLM Agents for Multimodal Neuroimaging Analysis and Research

Lujia Zhong, Yihao Xia et al.

NeuroAgent is an LLM-driven agentic framework designed to automate complex, multimodal neuroimaging analysis workflows, spanning preprocessing to downstream tasks. It utilizes a hierarchical multi-agent architecture with a feedback-driven Generate-Execute-Vali…

abstract pdf

NeuroAgent Framework Overview. The system comprises a Central Orchestrator (planning), Specialized Modality Agents (execution), and a Feedback-Driven “Generate-Execute-Validate” engine that enables reflective self-correction. A Human-in-the-Loop interface allows researchers to supervise and intervene at critical decision points. — NeuroAgent Framework Overview. The system comprises a Central Orchestrator (planning), Specialized Modality Agents (execution), and a Feedback-Driven “Generate-Execute-Validate” engine that enables re…

PACZero: PAC-Private Fine-Tuning of Language Models via Sign Quantization

Murat Bilgehan Ertan, Xiaochen Zhu et al.

PACZero introduces a novel, highly private fine-tuning method for language models based on **PAC (Probably Approximately Correct) Privacy**, specifically targeting resistance to Me…

abstract pdf

Recursive Agent Optimization

Apurva Gandhi, Satyaki Chakraborty et al.

Recursive Agent Optimization (RAO) is a reinforcement learning method designed to train agents capable of recursively spawning and delegating sub-tasks to new instances of themselv…

abstract pdf

SkillOS: Learning Skill Curation for Self-Evolving Agents

Siru Ouyang, Jun Yan et al.

SkillOS introduces a novel reinforcement learning (RL) framework for self-evolving agents to automatically curate a repository of reusable skills from experience. It pairs a frozen…

abstract pdf

StraTA: Incentivizing Agentic Reinforcement Learning with Strategic Trajectory Abstraction

Xiangyuan Xue, Yifan Zhou et al.

StraTA introduces an explicit, sampled trajectory-level strategy to agentic reinforcement learning, addressing the limitations of purely reactive LLM agents in long-horizon tasks. …

abstract pdf

Superintelligent Retrieval Agent: The Next Frontier of Information Retrieval

Zeyu Yang, Qi Ma et al.

The paper introduces the **Superintelligent Retrieval Agent (SIRA)**, which aims to overcome the limitations of iterative, exploratory retrieval by compressing multi-round searches…

abstract pdf

See all 20 papers →

§ III