Notas
Índice cronológico de todas as notas do Scholion.
5359 notas · última:
JSON is not JSON Across Languages | Dochia CLI Blog
JSON is simple on paper, but different languages and libraries interpret numbers, strings, dates, key order, and nulls differently, creating interoperability bugs across systems.
In Defense of C++
The post argues that C++’s complexity, age, and safety criticisms are overstated, and that readable code, modern features, and good practices matter more than the language’s reputation.
Will Vibe Coding Kill LowCode - DEV Community
The post argues that vibe coding will not replace LowCode soon because of cost, nondeterministic results, adoption inertia, security risks, technical debt, and user preference for graphical interfaces.
The thing is ... I love programming !
The author says AI and layoffs do not change his attachment to programming. He values code as an act of creation and expression, not as competition, speed, or recognition.
Skip to the end
The article argues that AI flips old software assumptions: developers should stop building tools that force users to model problems and start delivering the outcomes directly, such as organizing content, executing rules, and writing messages.
Trump Tariffs Cause Chaos on eBay as Every Hobby Becomes Logistical Minefield
The Trump administration’s changing tariff rules and the end of the de minimis exemption are making cross-border buying harder and more expensive, especially on eBay, while sellers and mail carriers restrict U.S. shipments.
The Biological Rulebook Was Just Rewritten—by Ants
A study on Iberian harvester ants says queens clone males from another species to keep workers and future mates available, while the issue also notes NASA exoplanet posters, Saturn’s new features, and a river navigability method.
Trump Take LEGO
LEGO has stopped shipping more than 2,500 individual bricks from its Pick a Brick program to the United States and Canada, with the de minimis exemption ending and tariffs making the service untenable.
Apanhado a Roubar (2025)
Crime/comédia negra de Darren Aronofsky (2025) com Austin Butler: ex-baseball, hoje barman em Nova York, se enrosca com a máfia ao cuidar do gato do vizinho.
The Design of the New Swiss Passport
The new Swiss passport, first issued in 2022 and designed by RETINAA, uses a visual narrative of landscapes and waterways, with UV-only security features that are meant to be both hard to counterfeit and part of the design.
LLMs are not like you and me—and never will be.
The post argues that LLMs are not human-like minds but autocomplete systems that lack proper world models, so they keep making obvious errors in time, history, economics, and factual reasoning.
How to write a great error message
The article argues that good error messages are clear, specific, actionable, and written in plain language. It criticizes vague alerts, unnecessary interruptions, and defaults that could be handled without modal dialogs.
IV SIVKF - Dia 03 - Por Thiago Silva - A melhor salada do mundo
Relato do terceiro dia de imersão, com caminhada cedo, cafés com convidados, prática informal, facas no fim do dia, Ginger Beer e uma defesa entusiasmada da salada do Little Greek.
IV SIVKF - Dia 01 por Thiago Silva: O Caramelo Machiatto da Vida
O texto usa chás, maror e caramelo machiatto para pensar a impermanência da vida, que alterna momentos doces e amargos sem mistura totalmente previsível.
How I Learn Technical Things Fast
The page lists eight habits the author says help him learn faster: prepare fundamentals, focus on one task, study in immersive settings, work on weaknesses, use active recall, seek feedback, think from first principles, and teach.
How to: Design for exception safety
Explains how exception safety depends on RAII, simple resource classes, and the three guarantees: no-fail, strong, and basic. It also lists class-design rules for constructors, destructors, and ownership.
I Vibe-Coded a Triton Kernel
The author describes trying to make GPT-OSS attention trainable by replacing a forward-only Triton kernel with a working backward pass, using Modal, ChatGPT, Claude Code, and Codex CLI to fix test failures until the harness passed.
Mastering the Decorator Pattern in PHP
The page shows how the Decorator Pattern avoids class explosion in PHP by wrapping a base object with small decorators that add features at runtime. It contrasts this with inheritance-heavy combinations and a large factory switch.
Forbes Under 30 Lister Ashley Edwards Attracts $1 Million in Venture Funding for MindRight, a Mobile-First Mental-Health Coaching Service
Ashley Edwards says MindRight Health makes mental-health coaching accessible by text for young people in underserved communities. The company raised more than $1 million in venture funding and works with partners including RWJBarnabas Health.
Silicon Valley’s Dream Tech Job Is Disappearing
Big tech jobs have lost the perks, openness, and security that once made them coveted. Layoffs, tighter control, and AI fears have made workers feel replaceable and more cautious.
Move fast and don't break (safety critical) things
Boom describes how in-house engineering software, software engineers embedded in hardware teams, and rapid iteration helped the small XB-1 team design and fly a supersonic jet quickly and with less budget.
The Art of Talking to AI: Prompts That Actually Work for Coding
The article says AI coding help works better when prompts are specific, include context, skill level, tools, output format, and requests for explanations or debugging help.
Why Your AI Never Works on the First Try
The article argues that AI output often needs many iterations because LLMs only move step by step through possibility space. That makes AI a cognitive exoskeleton that needs human navigation.
How o3 and Grok 4 Accidentally Vindicated Neurosymbolic AI
The essay argues that pure scaling has hit diminishing returns, while LLMs improve when paired with symbolic tools. It presents this as a belated vindication of neurosymbolic AI.
GPT-5: Overdue, overhyped and underwhelming. And that’s not the worst of it.
Gary Marcus argues that GPT-5’s debut exposed the limits of hype around OpenAI, with public disappointment, lingering errors, weak generalization, and no evidence that pure scaling gets to AGI.
Retries Have an Evil Twin: Duplicates
The post explains why at-least-once delivery leads to duplicate work and compares four application-level ways to stop it: database constraints, in-memory dedup, Redis, and broker support.
How to Implement Multitenancy in ASP.NET Core with EF Core
The post explains multitenancy with a discriminator column, then shows how EF Core query filters and SaveChanges can isolate tenant data, set TenantId automatically, and support login and header-based tenant selection.
How to Implement Audit Trail in ASP.NET Core with EF Core
The post shows how to add audit trails in an ASP.NET Core app with EF Core by using a separate AuditTrail entity, current-user lookup, and DbContext change tracking.
Notte: Reliable Browser AI agents
Notte is a full-stack framework for building and deploying web automation agents, combining deterministic scripting with AI to reduce costs, improve reliability, and support structured output, sessions, vaults, personas, and scraping.
The Best Way To Map Objects in .Net in 2024
The post compares manual object mapping with AutoMapper and Mapster in .NET, and argues that manual mapping with required properties is safer, clearer, faster, and easier to debug.
Logging Best Practices in ASP.NET Core
The post recommends Serilog for ASP.NET Core, using the right log levels, filters, structured logging, sensitive-data protection, error logging, log rotation, async sinks, and a centralized UI like Seq.
Reflections on the React community
Lee Robinson reflects on React’s growth, the difficulty of community management, and how commercial and non-commercial OSS create different incentives around feedback, RSC, and Next.js.
Notes to myself
A list of 65 reminders about systems, customer choice, leadership, trust, discomfort, and long-term work, framed as practical beliefs for making better decisions and building durable results.
How to Determine What is Important
The article argues that importance is clarified by aspirations, mortality, happiness, basic needs, and integrity, all of which help define the identity a person wants to build.
Pesquisa Mostra Que “Descansar” Pode Não Ser a Melhor Forma de Recuperar Energia
Um estudo diz que usar o tempo livre de modo intencional pode elevar bem-estar, energia, humor, propósito e desempenho, com menos risco de burnout do que lazer passivo.
We'd be Better Off with 9-bit Bytes
The post argues that if 9-bit bytes had become standard, many later limits would have been less painful: IPv4, UNIX time, Unicode, pointers, and other encodings would have had more room before exhausting.
Laurence Tratt: LLM Inflation
Tratt describes “LLM inflation” as using LLMs to turn short content into long prose and long prose into short summaries. He argues this can reward obfuscation and hide weak thinking.
We replaced passwords with something worse
The post argues that email-or-phone login codes weaken account security because users cannot tell whether a code prompt is legitimate, and password managers do not help against this phishing pattern.
New Method Is the Fastest Way To Find the Best Routes
A new shortest-path algorithm breaks the long-standing sorting barrier by avoiding full sorting of the frontier. It runs faster than Dijkstra’s on directed and undirected graphs.
The Rise of Silicon Valley’s Techno-Religion
The article portrays Lighthaven and the Rationalist-Effective Altruism network as a near-religious force in AI, shaping tech culture, funding, careers, and fears about humanity’s future.