Belisar Hoxholli

@jeffrey_way I think part of that is because the ease with which you can build everything, also creates a perception that whatever you are building is not that valuable. Focus and commitment are the hot commodity now, has been for a while.

Can you relate to this awkward tension I feel of being endlessly excited by what AI now unlocks (you can build anything you want), but with this constant underscore of depression that I can't explain?

@Grady_Booch Imagination.

It is increasingly clear that large language models - by the very nature of their architecture - are incapable of producing anything beyond the mediocrity of their training data. For me, the interesting question is this: why are humans able to do so?

New art project. Train and inference GPT in 243 lines of pure, dependency-free Python. This is the *full* algorithmic content of what is needed. Everything else is just for efficiency. I cannot simplify this any further. gist.github.com/karpathy/8627f…

all these "ai changes everything" threadboys look like they fuck with airpods on

fragments: the consequences of absolute skepticism, cognitive debt with AI, will AI remove fun of programming, the future of source code, how to survive AI contributions to open source, explaining transformers to 12/62-year-olds martinfowler.com/fragments/2026…

Software development quietly flipped from popularity by authority to authority by popularity. Most popular technical opinions now carry the same credibility as influencer-marketed VPNs and fitness supplements.

@nolannordlund Syntax becoming unpredictable, resistance to code quality and type safety changes, additions feeling half done. Growing pains mostly, but feels much better now.

@belisar What had you questioning the direction?

I was sceptical about the framework direction a couple of years back but all recent additions, from small things to bigger features, are helping it beat itself to excellence. Seems funding helped get over resource constraints and growing pains. Ripe for AI productivity gains too.

AI is giving us countless different ways to skin a cat on top of the countless previous ways we could skin a cat. The actual superpower might be to leave the cat alone.

This.

@iamdevloper Said developer's imposter syndrome will judge it, trash it, obsess over it, hide it away and eventually learn it's far better to let AI write it, so they're not the one who wrote it poorly.

Feeling philosophical today: If a developer writes a code and no one is around to judge it, is it still poorly written?

@Devinbuild Marketing -> Sales -> Coding. I have rarely seen someone with a good audience fail at monetizing it when they want to. I have seen graveyards with all sorts of great ideas and side projects.

What skill will get founders to $10k MRR 1. Coding 2. Sales 3. Marketing

@fchollet It's not changing that much at a fundamental level. New tools, more productivity, more opportunities, more parallel exploration. All that is true. We're living through weird times where hype has become its own industry, distorting the perception of reality.

Folks who work in AI or software engineering feel like the world is changing exponential fast. Because *their* world is changing exponentially fast. Folks in structural engineering or aeronautical engineering might not share the same sentiment.

@TheGeorgePu If I had a dollar for each thing that's supposedly been dead for the last 5 years and is still rolling, I'd be planning one of those fancy expedition cruises😅

Nasdaq dropped 2.3% today because Anthropic added enterprise plugins to Claude. The market finally sees it. SaaS is dead.

A bargained down design is much easier to spot than a bargained down codebase. A messy drawer is easier to tolerate than a messy table top. That said, designers sometimes optimize for the delight of the 2% rather than usefulness for the 98%. The 2% work inspires. The 98% is what pays the bills.

i think i am cooked as a dev way before designers models just do not want to make nice designs

@christophrumpel Depends. It's not different than coding for me. It becomes very clear if you have been getting what you get, or getting what you want. I would read something you source, curate, research, think, reason with AI gluing things. AI sourced ideas don't really result in good content.

Been experimenting with letting my bot help me with blog posts: ➡️ I talk to the bot about the idea, key takeaways, examples etc. ➡️ Bot creates a written version using the style and tone of my 100 posts ➡️ I fine tune wording etc. Would you still read it or no-go for you?

A modern airliner can land in CATIII conditions with zero visibility, keep the centreline and get people where they need to be after a cruise, climb and descend, at the right speed and efficiency. Is it conscious about what it’s doing? Nope. Is it needed for it to be conscious? Nope.

LLMs don't know what they're doing.

What would you like to measure that you cannot ask? If you can’t ask maybe it is an exec issue. Airline pilots and sales people are measured differently but have similar difference to the one that is puzzling you. You do measure development outcomes in system stability, error recovery, infrastructure cost, problem resolution, time to market capabilities, etc. It’s not that complicated.

developers are rare in that they have a job where they get to say "don't measure me" and somehow everyone just accepts it sales people get dashboards tracking every call, every email, every deal marketers get attribution on everything but suggest looking at how devs actually use their tools, and suddenly you're the villain we have access to more behavioral data than ever before... and we're too scared to use it :)

Customer email done right vs done wrong. @laravelphp: clicked, read, felt respected (great update!). @HarvardBiz: signed up, asked for essential emails only, got buried in spam, unsubscribed from the service entirely. Respect people's time.

@dillon_mulroy Oh, this is interesting and very much a breath of fresh air. Closest thing I've seen to what future tooling might look like when it catches up with the automation-first paradigm. Working on a similar direction as well. Looks great!

not too shabby

@jeffrey_way @lileshjadav @LaraconIN @abigailotwell You could have a cooking channel and I would still listen to recipes I have no intention of cooking. Being an exceptional educator, especially in complex subjects, takes talent and years of dedication to master.

@lileshjadav @LaraconIN @abigailotwell Ah that’s nice of him. I’m just attracted to really well built tools. 😘

Taylor otwell sharing his journey on stage building laravel. not being burned out & how greatful and humble he is & he thanked @jeffrey_way for building @laracasts. Thank you @taylorotwell for taking time visiting @LaraconIN & Speaking your heart.