Daniel Prada

Software engineer at @CBSSports that frequently debates how many exclamation points are socially permitted in a work message.

[JD Vance @JDVance]

Twenty years ago we invaded Iraq. The war killed many innocent Iraqis and Americans. It destroyed the oldest Christian populations in the world. It cost over $1 trillion, and turned Iraq into a satellite of Iran. It was an unforced disaster, and I pray that we learn its lessons.

[Daniel Prada @PradatiusD]

As a developer, this past month I've been submerged in a sea of X posts exposing various iterations of nihilistic predictions over the future of software development as coding agents multiply our productivity, yet put into open question our relevance in the future.

Software is instruction, and agents scale the amount of instructions by at least a multiple. Once it is common that agents spawn agents, we arrive at exponential instruction generation.

But the conversation conflates instruction generation with the hidden responsibility of maintaining a codebase.

Can AI agents scale judgement? Accountability? Trust? Can we see a world where it would be legally liable? To all these, it seems at least for me, the answer is no. Making software and knowledge will be cheaper than ever before, but maintaining it, expanding it, verifying it, validating it, will be the challenge. It's not useful to me if an AI agent builds an entire CRM if I cannot trust that it handles customer data securely, produces accurate analytics, handles concurrency for meaningful customer moments.

I'm sure the skill of the future will be on how to orchestrate agents to reduce slop generation.

In other domains I also see the same, analytics, medicine, law, health, sure AI can produce knowledge that sounds intelligent, correct, and trustworthy, but would you put your business' future, your family's estate, your child's health on something that is built and designed to please you, to give you a smart-sounding answer, even if it is knowledge that it does not have.

My take is that as companies flair how little coding their devs do, a silent technical debt will accumulate that is of a new species and of a greater magnitude. The junior devs being deprived of early jobs (as companies slash teams to glorify how great their AI tools are) will just have to be patient as those same companies will realize the AI knowledge and technical debt generated by these agentic coding tools leaves them blind, unsure, and unclear.

It has never been such a fantastic time to be a developer, and at the same time, it is a dizzying amount of change to see. Never will developers have to prove their value the most in time, but at the same token, never will their judgment become most valuable.

[Chase Passive Income @chasedownleads]

My friend has a secret hack to make $4,317 a month

And it’s so easy

Each morning, he shows up at a building

He calls it an “office”

Types on his computer for 8 hours

Maybe takes a short lunch break in the middle

Does this every day (except Saturday and Sunday)

And every two weeks, like clockwork, money mysteriously appears in his bank account

This blew my mind when he told me about it

Why isn’t everyone doing this?

[Supabase @supabase]

Handy

[Daniel Prada @PradatiusD]

https://t.co/hEW9XX1oHg

[Daniel Prada @PradatiusD]

NYT: A.I. Chatbots Defeated Doctors at Diagnosing Illness https://t.co/cbkJ5bvTRS

[Daniel Prada @PradatiusD]

https://t.co/e4zNMtRaoz

[Daniel Prada @PradatiusD]

https://t.co/duOyXhSRUb

[Memes Universidad Nacional @nacional_memes]

Te quiero mucho @BancoRepublica 🐶

[TONY @MisterIntruso]

Esto se salió de control 🤣🤣

[dennis hegstad @dennishegstad]

🤔🤔

[Emil Privér @emil_priver]

React 20 looking good!

[Daniel Prada @PradatiusD]

https://t.co/oWnD01UmOQ

[Daniel Prada @PradatiusD]

Article suggests "Take a deep breath and work on this problem step-by-step." in GPT4o prompts improves mathematical reasoning.

Probably works just as well on humans too 😅
https://t.co/Xrm971j6HL

[Daniel Prada @PradatiusD]

https://t.co/hlVFHlnAXu

[Daniel Prada @PradatiusD]

Google CEO Sundar Pichai on AI-powered search and the future of the web https://t.co/c44DfHKiYO via @Verge

[USA Ultimate @USAUltimate]

The final score that secured @BMoUltimate its fourth title in school history! 🏆

#USAUCollegeChamps | #USAUltimate

[Daniel Prada @PradatiusD]

Google I/O Summary:
https://t.co/9tivQjQVjV

[Daniel Prada @PradatiusD]

AI is going to be in every application that people will welcome insert chips into their brains to avoid seeing it 😂
https://t.co/uUAssbTyy1

[Daniel Prada @PradatiusD]

An AI for nervous new parents, An AI for low-level paralegal work, AI for conversations with the elderly. It just seems more realistic to me that the new generation will treat models like how I treated search sites, after we hit the limits of what a general solution provides.

[Daniel Prada @PradatiusD]

On a more serious note, thinking about AI, I do see the future as one where, like search, we first had a single search engine we went all to, whether it was AskJeeves, Yahoo, or now Google/Bing for any query. But over time, we saw the unbundling of that technology.

Thread 🧵

[Daniel Prada @PradatiusD]

A model for finding the correct legal precedence, a model for providing medical knowledge, an image recognition model for finding good deals online matching what you prefer, an AI tutor finding you the best content to learn, an AI code assistant, an AI historian.