Andrés Meza-Escallón

"If it hurts, do it more often" — Martin Fowler

These gaps cannot be eliminated; they can only be shrunk by failing faster and get feedback frequently

Evidence es louder than opinions — David Pereira #SoftwareDevelopment #UIUXdesign

We’ve found the root cause of AWS outage today. Sorry guys.

If this Karpathy interview doesn't pop the ai bubble, nothing will. 10 brutal quotes: 1. LLMs don’t work yet They don’t have enough intelligence, they’re not multimodal enough, they can’t use computers, and they don’t remember what you tell them. They’re cognitively lacking. It’ll take about a decade to work through all of that. 2. When you boot them up, they always start from zero They have no distillation phase, no process like sleep where what happened gets analyzed and written back into the weights. 3. What’s stored in their weights is only a hazy recollection of the internet It's just a compressed blur of 15 trillion tokens squeezed into a few billion parameters. Their context window is just short-term working memory. 4. They’re good at imitation, terrible at going off the data manifold Too much memory, not enough reasoning. We need to strip away the memorized knowledge and keep the cognitive core: the algorithms, the magic of intelligence, problem-solving, strategy. 5. We’ve probably recreated a cortical tissue, pattern-learning and general, but we’re still missing the rest of the brain No hippocampus for memory. No amygdala for instincts. No emotions or motivations. 6. They memorize perfectly but generalize poorly If you give them random numbers, they can recite them back. No human can do that. That’s the problem: humans forget just enough to be forced to find patterns. 7. Anything truly new, code that’s never been written before, ideas that have no template; they stumble They’re still autocomplete engines with perfect recall and no understanding. Until we find that cognitive core, intelligence stripped of memory but full of reasoning, they’ll stay brilliant mimics, not minds.

The tactical layer of development is becoming commoditized. But the strategic layer — that’s still very much a human game. — Craig Adam #ai #SoftwareDevelopment

Assumptions that aren’t based on well-established facts are the bane of all projects. —Andrew Hunt & David Thomas, The Pragmatic Programmer @SoftwareDevelopment #SoftwareEngineering

Splashdown confirmed! Congratulations to the entire SpaceX team on an exciting eleventh flight test of Starship!

Predictable patterns so AI can see what “right” looks like. Tight constraints so it doesn’t wander into dangerous territory. Curated examples what you show AI becomes what it repeats. Clean abstractions for token-hungry models trying to reason across files. — Craig Adam

You need to sometimes think much harder. Now there is human judgment, evidence and AI thinking. AI is a very good complementary tool that helps us with some of the deep thinking. — Itamar Gilad #SoftwareDevelopment #ProductDesign #UIUX

Always take small, deliberate steps, checking for feedback and adjusting before proceeding. Consider that the rate of feedback is your speed limit. You never take on a step or a task that’s “too big.” — Andrew Hunt & David Thomas, The Pragmatic Programmer

There are two aspects of time that are important to us: concurrency (things happening at the same time) and ordering (the relative positions of things in time). — Andrew Hunt & David Thomas The Pragmatic Programmer

Focus on solving current over future problems — David Pereira #SoftwareDevelopment

What’s a task that’s too big? Any task that requires “fortune telling” — Andrew Hunt & David Thomas The Pragmatic #Programmer #SoftwareDevelopment #SoftwareEngineering

#Agile planning depended, to a significant degree, on decomposing work into small enough pieces that we could complete our features within a single sprint, or iteration. — Dave Farley #SoftwareEngineering #SoftwareDevelopment

A thing is well designed if it adapts to the people who use it. — Andrew Hunt & David Thomas, The Pragmatic Programmer #SoftwareDevelopment #SoftwareEngineering

Instead of wasting effort designing for an uncertain future, you can always fall back on designing your code to be replaceable. — Andrew Hunt & David Thomas, The Pragmatic #Programmer #SoftwareDevelopment #SoftwareEngineering

Stop the nonsense. Don't build an AI Agent unless there's no other alternative. For most applications, you'll be better off: 1. With a simple rule-based system 2. Upgrade to static models when complexity increases 3. Implement simple, well-prompted LLM calls 4. Build an LLM-based workflow with predefined code paths Then, and only then, AI agents might be a good alternative. Every single option on the above list is easier to implement and more maintainable than an agent.

Sometimes a design just feels wrong, or some requirement makes you feel uneasy. Stop and analyze these feelings... Explore them. The chances are that there’s something lurking in that dark doorway. — Andy Hunt & Dave Thomas, The Pragmatic Programmer

Yeah, anyone can do #VibeCoding for throwaway stuff, but when things get serious...