HackerNews Digest

Jimi Hendrix’s 3 Feb 1967 “Purple Haze” session used a custom Octavia pedal (Roger Mayer) alongside a Fuzz Face, wah‑wah, Marshall 100‑W amp and studio acoustics. The author reproduced the full analog signal chain in ngspice, modeling guitar pickups (6 kΩ, 2.5 H) and cable capacitance, and simulated both germanium and silicon Fuzz Face variants. Key findings: • Fuzz Face is a low‑impedance (≈20 kΩ) two‑transistor feedback amp that turns sinusoidal input into near‑square‑wave fuzz; reducing guitar volume restores a sinusoid (“cleanup effect”). • Octavia’s rectifier flips waveform troughs, effectively doubling frequency content and adding a bright octave‑up harmonic. • Wah‑wah functions as a sweepable band‑pass filter (≈300 Hz → 2 kHz), producing vowel‑like articulations. • Uni‑Vibe cascades four phase‑shift sections modulated by photoresistors, adding low‑frequency motion. • Driving the Marshall near saturation and exploiting room‑coupled acoustic feedback extends sustain and creates controllable oscillation modes. All circuit schematics, ngspice files, and Python scripts are publicly available on GitHub (nahorov/Hendrix-Systems-Lab). The work reframes Hendrix’s tone as an engineered, reproducible system.

A lawsuit filed on February 23 in the U.S. District Court for the Southern District of New York (Case 1:26‑cv‑1504) accuses Jane Street Group LLC of insider trading and market manipulation related to the May 2022 TerraUSD (UST) collapse that erased about $40 billion in value. The complaint, brought by Todd R. Snyder, the court‑appointed administrator of Terraform Labs, alleges that after Terraform withdrew $150 million of UST from the Curve 3‑pool without public notice, a wallet linked to Jane Street withdrew an additional $85 million within ten minutes—a move described as Jane Street’s largest single swap and a catalyst for the loss of confidence in UST. The suit claims Jane Street obtained material non‑public information via a back‑channel created by former Terraform intern Bryce Pratt, who allegedly communicated with Terraform’s business‑development lead. Co‑founder Robert Granieri and employee Michael Huang are also named. Jane Street denies the allegations, attributing losses to fraud by Terraform’s management (Do Kwon). The case follows a related suit against Jump Trading alleging similar conduct. If the claims proceed, Jane Street may be compelled to produce internal communications and trading data concerning the TerraUSD market.

The page titled “http://info.cern.ch” identifies itself as the home of the first website, offering a brief navigation prompt “From here you can:” without further content.

Contemporary glass artist Simon Berger creates sculptural works by striking glass panes with a hammer, using the resulting cracks to explore material depth, transparency, and contrast. The glass serves both as structural support and as a visual “handwriting,” with rapid, brief blows producing stronger tonal variations. Berger’s process treats the hammer as an effect‑amplifying tool rather than a destructive instrument. Originally a carpenter, he began with spray‑painted portraits, then expanded to wood and mechanical assemblages, including work on used car bodies. The concept of using automotive windshields as a medium emerged from this experience. Berger cites a fascination with human faces, noting that safety glass transforms abstract fogging into recognizable figurative forms that engage viewers. His practice merges a background in carpentry and mechanics with a distinctive glass‑cracking technique to generate abstract‑figurative visual effects.

- HP’s CFO reported that RAM’s share of a PC’s bill of materials has risen from roughly 15‑18 % in fiscal Q4 2025 to about 35 % for the remainder of the fiscal year, reflecting a severe memory shortage. - Memory costs have increased approximately 100 % sequentially, with further price escalation expected throughout fiscal 2026 and likely into fiscal 2027. - The Personal Systems market is projected to contract by double‑digit percentages this calendar year as higher component prices suppress consumer demand. - HP anticipates the most pronounced financial impact from the RAM shortage in the second half of its fiscal year, citing rising DRAM and NAND prices as primary drivers of input‑cost volatility. - To offset higher RAM expenses, HP has raised PC prices while noting that roughly one‑third of its Personal Systems margin derives from non‑RAM categories such as IT services and peripherals.

- MCP agents preload the full JSON‑Schema for every available tool (e.g., 84 tools across 6 servers), incurring a large token cost at session start (≈600 tokens for tool discovery plus additional tokens for each call). - A CLI‑based approach (generated with CLIHub) lists only tool names and locations; detailed schemas are fetched lazily when the agent invokes a command, reducing discovery tokens to a few tokens per tool. - Benchmarks show the CLI method uses about 94 % fewer tokens overall compared with MCP’s eager loading. - Anthropic’s “Tool Search” loads a searchable index and fetches full schemas on demand, cutting token usage by ~85 % but still incurs the full schema cost per tool and is limited to Anthropic models. - CLIHub provides an open‑source converter that creates CLIs from MCP definitions in a single command, preserving OAuth and API compatibility while achieving the token savings across any LLM.

Notepad (v 11.2512.10.0) receives several enhancements for Windows 11 Insiders in the Canary and Dev channels. Markdown support is expanded to include strikethrough and nested‑list syntax, accessible via the formatting toolbar, shortcuts, or direct markup. A new welcome dialog introduces core features and can be reopened via a megaphone icon. AI‑assisted “Write,” “Rewrite,” and “Summarize” functions now stream results, showing partial output sooner; these features require a Microsoft‑account sign‑in. Feedback should be filed in Feedback Hub under Apps → Notepad. Paint (v 11.2512.191.0) adds two features. “Coloring book,” an AI‑driven tool in the Copilot menu, generates custom coloring‑page images from text prompts; it is limited to Copilot+ devices and also requires a Microsoft‑account sign‑in. A fill‑tolerance slider for the Fill tool lets users adjust precision, enabling cleaner fills or artistic effects. Feedback for Paint is to be submitted via Feedback Hub under Apps → Paint.

Zyora-Dev’s GitHub repository “zse” provides the Zyora Server Inference Engine for large language models (LLM). The project is packaged on PyPI, indicating it can be installed via the Python package index. Compatibility requires Python 3.11 or newer. Licensing information is displayed, though the specific license type is not detailed in the extracted text. Deployment options are highlighted with badges for Railway and Render, suggesting that the engine can be hosted on either platform. An interface message reads “You can’t perform that action at this time,” implying a temporary restriction on certain repository actions. The available visual cues consist of alt‑text labels for the PyPI badge, Python version requirement, license indicator, and the two deployment platforms.

American bus routes typically place stops every 200–250 m (≈5–8 stops per mile), far closer than the 300–450 m spacing common in Europe. Frequent stopping adds dwell, deceleration and acceleration time, slowing buses to about 8 mph in cities such as New York and San Francisco, raising labor costs (≈70 % of operating budgets) and limiting investment in stop amenities. “Stop balancing” – increasing stop spacing to roughly 400–600 m (≈1,300 ft) – can be implemented cheaply by removing signs and revising schedules. Pilot projects show measurable gains: San Francisco saw 4.4‑14 % higher speeds; Vancouver saved ≈$500 k annually and cut travel time by ~5 min; Portland achieved a 6 % speed rise with a 90‑ft spacing increase; Los Angeles’ limited‑stop service boosted speeds 29 % and ridership 33 %; Washington DC recorded 22‑26 % speed gains. Modeling indicates coverage loss is modest (1‑13 %). Faster, more reliable service reduces required vehicles and labor, frees resources for better shelters, and can increase overall network accessibility without extensive new infrastructure.

The Respectify analysis of a snippet titled “Respectify – Improve Online Discourse” reports no logical fallacies but flags two issues: the phrase “dumb animals” is labeled objectionable for disrespecting bears and lacking supporting evidence, and the sentence “every bear is like this. Polar bears and grizzlies, they all think the same thing.” is identified as a negative‑tone overgeneralization that attributes human thoughts to animals. The content is marked as low effort and receives an overall quality score of 1. A brief note indicates that the full API response contains additional information beyond this summary.