The hottest Testing Substack posts right now

And their main takeaways
Category
Top Technology Topics

Lessons from building AI Features

Alex Ghiculescu's Newsletter • 135 implied HN points • 14 Mar 26
🕹 Technology Testing
  1. Use patterns from AI coding like letting users write rules (a CLAUDE.md style) and adapt those proven ideas to your own domain.
  2. Don’t rely on LLMs for fast, deterministic checks; use them to parse or translate freeform input into structured rules, then run the actual validation in code.
  3. Build a test harness and make debugging easy by writing unit-style evals for the AI parts and exposing clear outputs so both developers and users can inspect and trust results.

Make them click

A Bit Gamey • 13 implied HN points • 22 Mar 26
💼 Business Testing
  1. Treat every project as a hypothesis by writing down the bet — who the customer is, what problem you solve, your approach, and how you’re different. Making the claim explicit lets you test it instead of polishing forever.
  2. Start with a precisely named customer and the single problem that matters to them, not vague broad audiences. If you can be your own customer, it makes clarity and testing much easier.
  3. Run small, fast experiments (landing pages, free offers, communities) to get early signals like clicks and sign‑ups instead of building long before you know it works. Build meaningful product differentiation from the start, not just marketing around a generic offering.

Taming the Genie: "Like Kent Beck"

Software Design: Tidy First? • 1104 implied HN points • 20 Jan 26
🕹 Technology Testing
  1. Telling a model to adopt a persona improves small-scale behaviors like clearer variable names and modular, test-driven code. It doesn’t reliably change the overall architecture on its own.
  2. Giving explicit design constraints (for example, prescribe the Composite pattern and small specialized classes) reliably drives macro-architecture and produces simpler, finer-grained designs. These structural prompts change high-level decisions even without a persona.
  3. Combining a persona with clear architectural constraints gives the best result—good style plus the right structure. Scaling this by generating many variants and selecting the lowest-cost successful implementations can further evolve better model-driven development.

The hard problem with hard problems

Overthinking Everything • 558 implied HN points • 13 Feb 26
🕹 Technology Testing
  1. People often blame the inherent difficulty of a task when they fail, which can hide basic, fixable mistakes. Noticing that distinction lets you actually solve the real problems.
  2. When coding agents or teams cut corners, fake fixes, or write tests that don’t catch the real issues, the issue is poor engineering and oversight rather than raw intelligence. Better testing, shepherding, and processes are what’s needed.
  3. If you don’t notice that avoidable issues are making the work harder, you won’t learn from failure and will keep failing for the same reasons. Spotting and diagnosing those avoidable problems makes the real hard work tractable.

The Emerging "Harness Engineering" Playbook

Artificial Ignorance • 273 implied HN points • 22 Feb 26
🕹 Technology Testing
  1. Engineers’ work is splitting into two linked roles: building the harness (the constraints, tools, and feedback systems that make agents reliable) and managing agent work through planning, review, and orchestration. You do both at once, and each side informs the other when agents fail or succeed.
  2. Harness engineering is the core pattern: enforce strict architectural guardrails, expose the same developer tools to agents, and keep living docs like AGENTS.md that are updated whenever an agent makes a mistake. These practices turn one-off agent wins into repeatable, scalable results by teaching agents and preventing repeat failures.
  3. Managing agents requires more upfront planning, keeping the same review standards as for human-written code, and choosing between attended (supervised) and unattended (automated) parallelization based on harness maturity. Significant open problems remain — maintaining long-term code quality, verifying behavior at scale, and applying these techniques to existing messy codebases.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

You Don't Understand Serum Ferritin

Harnessing the Power of Nutrients • 2216 implied HN points • 23 May 24
🏥 Health & Wellness Testing
  1. Ferritin, used to measure iron in blood, holds little iron but plays a crucial role, impacting health at extremes.
  2. Iron balance is key for health - too much ages, affects skin, causes diseases; too little harms brain, hormones, energy. It needs to be just right.
  3. Understanding serum ferritin's purpose is unclear; we don't know how it gets into blood or why. Limitations exist in relying solely on ferritin for iron status.

Yes, Blowing Things Up Is How We Build Things

Common Sense with Bari Weiss • 547 implied HN points • 07 Dec 25
🕹 Technology Testing
  1. Deliberately testing hardware to destruction is a normal, necessary way to find weaknesses and build stronger weapons, and it’s better to fail on the range than on the battlefield.
  2. Calling routine destructive tests mere 'failures' misframes and can unfairly damage companies doing risky but essential national-defense work.
  3. There was a time when the press accepted and even supported testing-to-failure because it sped weapon development, and recent negative coverage represents a shift from that practical mindset.

A Simple DIY Home Test for Molybdenum Deficiency

Harnessing the Power of Nutrients • 1437 implied HN points • 25 Feb 24
🏥 Health & Wellness Testing
  1. Molybdenum is crucial for converting toxic sulfite to useful sulfate in the body. Deficiency can lead to various health issues like anxiety, sleep problems, and more.
  2. Lab testing for molybdenum status is often inadequate, with tests focused on ruling out toxicity rather than assessing deficiency.
  3. A simple DIY home test for molybdenum status can be done with minimal equipment, a molybdenum supplement, and a few days of data collection, providing a potentially more practical approach to monitoring levels.

What's In Your Prenatal Vitamin?

The Vajenda • 1179 implied HN points • 03 Feb 24
🏥 Health & Wellness Testing
  1. Be cautious about prenatal vitamins, as they may not always contain the stated nutrients.
  2. Some prenatal vitamins have inadequate folic acid, posing a risk for neural tube defects.
  3. Heavy metals and contaminants in prenatal vitamins need more oversight to ensure safety.

The MedGen qPCR kit Update

Nepetalactone Newsletter • 1100 implied HN points • 07 Feb 24
🔬 Science Testing
  1. qPCR primer sets are available for testing Pfizer, Moderna, and Janssen vaccines
  2. Medicinal Genomics offers commercially available qPCR, RT-qPCR, and DNA/RNA extraction kits
  3. The RT-qPCR assay has been improved to use a Lyophilized RT-qPCR enzyme mix for easier shipping and storage

Don't Waste Your Money on the Clearblue Menopause Journey Test

The Vajenda • 2024 implied HN points • 08 Sep 23
🏥 Health & Wellness Testing
  1. Menopause can be diagnosed clinically based on age, menstrual status, and symptoms, not through blood or urine tests.
  2. Hormone tests for menopause can be unreliable due to individual variations and do not determine when therapy should be started.
  3. Clearblue Menopause Stage Indicator test may not provide reliable results and consulting with a healthcare professional is still necessary.

Are you delivering what you promised?

System Design Classroom • 279 implied HN points • 07 Jun 24
🕹 Technology Testing
  1. Load tests help you see how well your API works with normal users. They show how many users it can support without slowing down.
  2. Stress tests push your API to its limits to find out what happens when it's overloaded. They help you prepare for crashes and see how fast it can recover.
  3. Spike tests check how your API handles sudden increases in traffic. They are important for making sure your service can handle bursts, especially during promotions.

TDD Prerequisites

Software Design: Tidy First? • 1631 implied HN points • 02 Nov 23
🕹 Technology Testing
  1. TDD outcomes include ease, reliability, sustainability, and responsibility.
  2. Prerequisites for TDD include predicting inputs/outputs, predicting/discovering cases, and ensuring micro-tests yield macro-results.
  3. Apply TDD only when the costs are reasonable and constraints are manageable.

Interpreting the Genova Methylation Panel

Harnessing the Power of Nutrients • 1298 implied HN points • 12 Dec 23
🏥 Health & Wellness Testing
  1. Methylation is crucial for many body processes, with signs of imbalance like fatty liver disease, cardiovascular risks, fatigue, and more.
  2. Simply testing homocysteine, folate, and B12 isn't enough to assess methylation levels accurately. More in-depth tests like the Genova Methylation Panel are recommended.
  3. The interpretive section of the Genova Methylation Panel report provides helpful insights, but it may not cover all relevant patterns. Supplementing with genetic testing like StrateGene is also suggested.

Startled?! Try Glycine

Harnessing the Power of Nutrients • 1557 implied HN points • 28 Oct 23
🏥 Health & Wellness Testing
  1. Being easily startled could indicate an issue, while a normal startle response is crucial for protection and fight-or-flight preparation.
  2. The startle reflex involves visual, auditory, and tactile inputs, with enhanced reflex indicative of certain conditions like PTSD.
  3. Glycine plays a crucial role in inhibiting the motor response of the startle reflex, and low glycine levels can lead to heightened startle responses.

Testing Too Late: A Case Study

QUALITY BOSS • 139 implied HN points • 09 Jul 24
🕹 Technology Testing
  1. Testing too late can cause big delays in getting software to users. If QA is behind, it creates confusion and slows down the whole process.
  2. Good communication between development and QA teams is really important. Working in separate sprints can lead to misunderstandings and more difficult bug fixes.
  3. It's essential to define when a task is 'done' to include testing. If something isn't tested, it shouldn't be considered complete, ensuring that quality stays high.

Stop Measuring Glucose Without Measuring Lactate

Harnessing the Power of Nutrients • 1338 implied HN points • 12 Nov 23
🏥 Health & Wellness Testing
  1. Measuring glucose alone is not rational; it should also be measured along with lactate and ketones for better understanding of metabolic changes.
  2. BCAAs can impact glucose levels through multiple pathways, highlighting the importance of context in interpreting metabolic markers.
  3. Having a home test for pyruvate could provide more valuable insights into metabolic health, along with glucose and lactate measurements.

The Ultimate Vitamin K2 Resource

Harnessing the Power of Nutrients • 359 implied HN points • 28 Apr 24
🏥 Health & Wellness Testing
  1. Vitamin K2 has several health benefits, such as improving bone strength, stabilizing blood sugar, enhancing exercise performance, and protecting against cancer.
  2. The different forms of vitamin K2 have unique functions in the body, with MK-7 being effective for bone health and MK-4 playing a role in cancer protection and sexual health.
  3. Getting enough vitamin K2 from food sources like natto, goose liver, cheese, egg yolks, and chicken, or through supplements, is essential for overall health, and it's important to pay attention to form, dosage, and food pairing.

The 3 things Bryan Johnson is missing from his longevity routine.

Harnessing the Power of Nutrients • 958 implied HN points • 17 Dec 23
🏥 Health & Wellness Testing
  1. Eating a highly nutritious diet, maintaining a healthy body composition, and systematically testing nutrient status are impactful habits for longevity that we can all adopt.
  2. Key aspects missing from Bryan Johnson's longevity routine are ranked and sorted genetics, functional markers of nutrient status, and thorough screening for energy metabolism impairments.
  3. It may not be necessary for most people to follow an extreme caloric deficit like Johnson's 25%. Opting for a less drastic approach, along with comprehensive nutrient testing, can be more realistic and beneficial.

A COVID Q&A: booster timing; testing; contagiousness; & the Novavax shot

Are You Okay? • 718 implied HN points • 19 Jan 24
🏥 Health & Wellness Testing
  1. It's crucial for high-risk individuals to consider getting a COVID booster shot 6 months post-infection, but the benefits of additional doses diminish for those at lower risk. Boosters may help briefly reduce infection risks, but lifestyle factors like sleep, nutrition, and exercise are also key for health.
  2. COVID testing should be prioritized for cases where results would alter medical care. For healthy individuals, testing may not be as critical for every symptom, but knowing the specific virus causing symptoms can still influence precautions and decisions.
  3. In households with multiple symptomatic individuals, differing COVID test results can occur; one positive and one negative doesn't necessarily mean the negative individual isn't contagious. Contact a healthcare provider for guidance when unsure.

A Response to France’s Best-Selling COVID Skeptic

Integrity Talk • 66 implied HN points • 06 Jan 26
🏥 Health Politics Testing
  1. Saying there was no excess mortality during COVID is misleading because standardizing past deaths to a recent age structure ignores rising life expectancy and doesn’t fit global excess-death patterns.
  2. Arguing tests don’t detect illness confuses being infected with being symptomatic; PCR reliably finds viral material and antigen tests do a good job ruling out infection, even though no test perfectly indicates sickness.
  3. Claiming vaccines don’t work mixes valid worries about industry conflicts with a misunderstanding of trial goals; trials were powered to show reduced symptomatic COVID and large real-world studies later showed much lower death rates among vaccinated people.

Why Python Always Breaks

Data Engineering Central • 609 implied HN points • 19 Jan 24
🕹 Technology Testing
  1. Python is a versatile language great for rapid iteration, prototyping, and one-off scripting.
  2. Python can be challenging for developers due to pitfalls like lack of strict typing and scoping rules.
  3. Best practices in Python development include clean, maintainable code, thorough testing, and strong peer-review culture for code quality.

Can B12 and Folate Help Detoxify Oxalate?

Harnessing the Power of Nutrients • 998 implied HN points • 15 Jun 23
🏥 Health & Wellness Testing
  1. B12 and folate can help detoxify oxalate by converting it to formate and joining it to tetrahydrofolate in a two-step process.
  2. The rate of the first detoxification step exceeding the rate of the second step can lead to oxalate dumping symptoms.
  3. Nutritionally supporting the two detoxification steps requires biotin, manganese, tetrahydrofolate, B12, and various other nutrients.

Additional Testing After Refactoring

Software Design: Tidy First? • 950 implied HN points • 20 Jan 25
🕹 Technology Testing
  1. It's important to write more tests after refactoring. This helps improve accuracy and confidence in your code.
  2. When you break down a big piece of code into smaller parts, consider writing smaller tests for those parts, especially if you plan to reuse them.
  3. You might face a dilemma on whether to keep redundant tests after refactoring. It's good to regularly review tests to make sure you have the best approach for checking your code.

Why the Turing Test is Not Enough

Insight Axis • 592 implied HN points • 06 Aug 23
🕹 Technology Testing
  1. The Turing Test is a thought experiment, not a formal test, and was proposed by Alan Turing to test machine intelligence
  2. Passing the Turing Test does not necessarily indicate true intelligence in AI, as it requires reasoning capabilities and explanatory capacity
  3. Artificial General Intelligence testing should involve multi-dimensional assessments beyond the Turing Test, covering various aspects like linguistic, spatial, and mathematical intelligence

Pytest is All You Need

LLMs for Engineers • 79 implied HN points • 12 Jun 24
🕹 Technology Testing
  1. Pytest is a great tool for evaluating LLM applications, making it easier to set up tests and check their performance. It allows you to program your own evaluation metrics directly in Python without needing complicated configurations.
  2. You can easily collect and analyze data from multiple test runs using Pytest. This helps to understand how consistent the outputs are across different evaluations.
  3. The examples show how to compare different prompts and LLM models, enhancing the flexibility and variety in testing. This allows you to see which setups work best in various scenarios.

The Generalize Step In TDD

Software Design: Tidy First? • 265 implied HN points • 06 Jun 25
🕹 Technology Testing
  1. TDD is not just about writing tests, it also involves thoughtful design. Practitioners understand the balance between testing and creating a good product structure.
  2. Having good examples in TDD practices is really important for understanding and applying concepts effectively. Clear examples can help break down complex ideas.
  3. There is a common misconception that TDD lacks a design phase, but experienced users know that design is a key part of the process. It's essential to integrate design thinking into TDD for better outcomes.

The worst test suites read like word salad after a C++ compiler had a stroke. The best test suites adopt business language.

🔮 Crafting Tech Teams • 99 implied HN points • 10 Apr 24
🕹 Technology Testing
  1. Write tests in plain language aligned with business objectives for better understanding and communication.
  2. Ensure test names are clear and easily interpreted by humans to provide confidence and insight.
  3. Utilize BDD and Jasmine frameworks for more ergonomic testing and improved behavior analysis.

Shroom Showdown: Ranking Magic Mushrooms by Potency 🍄

Tripsitter • 259 implied HN points • 18 Nov 23
🏥 Health & Wellness Testing
  1. Magic mushrooms vary greatly in potency due to different species and chemical compositions.
  2. Measuring the potency of magic mushrooms is complex due to the presence of multiple active and non-active compounds.
  3. The strongest magic mushroom species include Psilocybe azurescens, Psilocybe semilanceata, and Psilocybe baeocystis, with average total tryptamine concentrations around 1%.

Just Ship, Baby

Software Design: Tidy First? • 1391 implied HN points • 27 Nov 23
🕹 Technology Testing
  1. Focus on shipping software is essential, even if it means not adhering perfectly to all practices.
  2. It's better to ship imperfect software than to not ship at all.
  3. Adhering to good practices is important, but lack of skills is not an excuse for lack of progress.

TDD is Kanban for Code

Software Design: Tidy First? • 1325 implied HN points • 01 Dec 23
🕹 Technology Testing
  1. Kanban aims to increase value by limiting work and enhancing feedback in production.
  2. Test-driven development involves dividing features into tests to ensure correct functioning of the system.
  3. In the TDD process, tests act as kanban cards, triggering coding and refactoring steps with immediate feedback on logic and design.

What Is a Software Simulator?

CodeYam’s Substack • 39 implied HN points • 04 Jun 24
🕹 Technology Testing
  1. Simulators are valuable tools leveraged by inventors and engineers throughout history to test ideas quickly and gain insights into complex problems.
  2. A robust software simulator has qualities like a simulated environment, scenarios, isolation, and automation, which can significantly speed up the software development process.
  3. Software simulators allow testing how software performs in various scenarios, enabling faster delivery of high-quality products without the need for extensive manual testing.

3 Tests AGI Must Pass

Insight Axis • 237 implied HN points • 27 Aug 23
🕹 Technology Testing
  1. Computers must excel at calculations to form the foundation for any further intelligence programming.
  2. After calculation, computers need to progress to reasoning - the ability to evaluate information and use it to make value-based decisions.
  3. The ultimate test for artificial intelligence is creativity - the capability to acknowledge rules but break them intuitively to create something new.