Published on

Daily Tech News - 2026-02-08

Authors

From high-stakes AI rivalries to the intricate mathematics of a sandwich, today’s tech landscape proves that even the most mundane systems hide surprising complexity.

The AI industry is bracing for a cultural collision as Anthropic prepares its first Super Bowl ad, directly challenging OpenAI’s shift toward monetization and advertising. While the giants debate regulation on the national stage, developers are focusing on the "human scale" of software. The emerging philosophy of "person-sized" microservices suggests that the most efficient codebases are those small enough for a single developer to hold entirely in their head. This shift toward manageable context is being bolstered by AI agents that now utilize ClauDE.md and JSONL session logs to maintain project awareness, and even automate the generation of brand-compliant corporate reports through simple Node scripts.

Complexity, however, isn't limited to code. A deep dive into gas station "forecourts" reveals the hidden networking and fluid dynamics required to sell fuel, while a satirical new "Sandwich Bill of Materials" (SBOM 1.0) reminds us that even a simple lunch can have a dependency tree of up to 47 ingredients. On the computational front, recent benchmarks of Fibonacci sequences highlight the eternal trade-offs between space and time, demonstrating how implementation choices—like Python’s native arithmetic versus the mpmath library—can radically alter performance.

We also see the friction between the past and the future: legacy tools like the original vi are increasingly incompatible with modern web security layers, and standalone video titlers—once the kings of broadcast text—have been relegated to history by desktop software. Yet, amidst the digital noise, there is room for the tangible, as the community celebrates the 2026 Kākāpō breeding season with artisanal craft, proving that even in a world of high-speed algorithms, some things remain worth savoring.

computing-large-fibonacci-numbers-4187db39

The previous post discussed two ways to compute the nth Fibonacci number. The relative efficiency of each algorithm depends on how it is implemented. In this post I will compare using Python’s integer arithmetic and the mpmath library.

  • Keywords: compute fibonacci, fibonacci number, fibonacci numbers, nth fibonacci, discussed fibonacci, fibonacci, digits fibonacci, 000th fibonacci, says fibonacci, 60 fibonacci
  • Source: johndcook.com

every-man-a-microservice-05316835

The right way to run a system is to have it made up of person-sized services. A single developer, working on that codebase full-time, can keep the whole codebase in their head. It enables a ton of offline mulling.

  • Keywords: organization designs, microservice organizations, organizations design, organizations systems, organization built, organization landscape, effective organization, structures organizations, organization, organizations
  • Source: grantslatton.com

fibonacci-numbers-and-time-space-tradeoffs-9aaba21a

A few days ago I wrote about Fibonacci numbers and certificates. This time I want to use the numbers to illustrate space tradeoffs. For small n, it’s fast and easy to compute the numbers up to n sequentially using the definition of the sequence. For large enough n, the former requires (extended) integer arithmetic, and the latter requires ( extended) floating point arithmetic.

  • Keywords: fibonacci certificates, certify fibonacci, number fibonacci, 000th fibonacci, post fibonacci, fibonacci number, large fibonacci, fibonacci numbers, millionth fibonacci, compute fibonacci
  • Source: johndcook.com

forecourt-networking-8e72686a

The front part of a gas station, where the pumps are and where you pull up your car, is called the Forecourt. The practicalities of selling gasoline, namely that it is a liquid sold by volume, make the forecourt more complex than you might realize. We might forget that, well, handling a fluid takes a lot of equipment as well... we just don't think about it.

  • Keywords: service gas, serve gasoline, forecourt infrastructure, forecourt networking, gas stations, gas station, gasoline stations, selling gasoline, states gas, station gas
  • Source: computer.rip

how-to-generate-good-looking-reports-with-claude-code-cowork-or-codex-ed8d6e10

How to generate good looking reports with Claude Code, Cowork or Codex. Every organisation has brand guidelines that nobody follows for internal documents. Reports go out in whatever template someone found on Google Docs, slides are a mix of three different colour schemes. I've replaced all of that with three prompts and a 50-line Node script.

  • Keywords: brand document, brand pdfs, templates report, report formats, extract design, brand design, create design, powerpoint template, format report, report slides
  • Source: martinalderson.com

k-k-p-mug-by-karen-james-e6b0b8f3

Kākāpō mug made by Karen James. It has four chicks and rimu fruit. I love it so much. It's in celebration of the 2026 breeding season.

  • Keywords: kākāpō mug, kākāpō chicks, kākāpō, kākāpō kākāpō, charismatic kākāpō, james kākāpō, 2026 breeding, strongdm ai, rimu fruit, pypi using
  • Source: simonwillison.net

sandwich-bill-of-materials-649a786c

SBOM 1.0 (Sandwich Bill of Materials) provides a machine-readable format for declaring the full dependency tree of a sandwich. A typical sandwich contains between 6 and 47 direct dependencies, each pulling in its own transitive ingredients. Each sandwich component MUST include the following fields:

  • Keywords: sandwich sbom, sbom sandwich, sandwiches dependency, sandwich reproducible, reproducible sandwich, sandwich infrastructure, sandwich restriction, sandwich ingredients, sandwich vulnerability, specification sbom
  • Source: nesbitt.io

self-improving-claude-md-files-36161fe7

ClauDE.md or AGENTS.md files give the agent context on the project. With Claude Code, these sessions are stored in ~/.claude/projects, with Codex storing them in ~/.codex/sessions. These agent logs are JSONL files which contain everything that happened in the agent session.

  • Keywords: agentic tools, agent build, agent session, improve agentic, agent search, agents md, agent logs, use agent, claude projects, file claude
  • Source: martinalderson.com

super-bowl-matchup-anthropic-vs-openai-bbf7110f

Anthropic is running its first Super Bowl ad, lampooning OpenAI’s recent decision to sell ads in ChatGPT searches. OpenAI gave lip service to AI regulation at the US Senate; Anthropic actually supports regulation.

  • Keywords: vs openai, differences openai, vs pepsi, bowl ad, openai, ad ads, ads, ads super, openai don, anthropic vs
  • Source: garymarcus.substack.com

the-original-vi-is-a-product-of-its-time-and-its-time-has-passed-e997a438

If you're reading this page because you've attempted to access some part of my blog (WanderingThoughts) or CSpace, you're using a browser (or client library) that my anti-crawler precautions consider suspicious. Your browser or client is advertising one of the following Accept-Encoding values: identity, gzip.

  • Keywords: encoding gzip, gzip ongoing, deflect gzip, gzip, gzip bombs, uncompressed replies, accept encoding, completely uncompressed, uncompressed, anti crawler
  • Source: utcc.utoronto.ca

the-scriptovision-super-micro-script-video-titler-is-almost-a-home-computer-076a6f76

Video titlers (also known as character generators) are devices that stamp bitmap data, usually text, on top of a video signal. Today's titlers are usually add-on boards controlled by a standard desktop computer, but for much of their existence they came as standalone devices with their own CPUs and video hardware.

  • Keywords: video hardware, consoles, potentially hardware, video chip, game consoles, video input, hardware, video interface, video signal, console built
  • Source: oldvcr.blogspot.com