Alex Vaystikh

If it ain't broke, break it

[Alex Vaystikh @bornio]

@seankenmig @AdamHoltererer Anywhere you need top performance on these benchmarks , for example

[Alex Vaystikh @bornio]

@AdamHoltererer great question!
depends so much on the performance of the model in the thread.

any data to proxy productivity? like, did the thread have a happy ending in the form a PR?

something like, number of prompts to get to PR, per model

[Alex Vaystikh @bornio]

@AdamHoltererer thanks!

yeah, i agree, this isn't showing what i was hoping it would.

i wonder if the population should be only people who have 2+ models, and to compete them against each other...

going to think about it some more 🤔

[Alex Vaystikh @bornio]

@karrisaarinen @linear all products are now a harness

some are better than others (like this one)

[Alex Vaystikh @bornio]

@iruletheworldmo High expectations after losing streak

[Alex Vaystikh @bornio]

@hikelikeme

[Alex Vaystikh @bornio]

@gabriel1 two words: jensen huang

[Alex Vaystikh @bornio]

@AdamHoltererer Thanks Adam!

[Alex Vaystikh @bornio]

@AdamHoltererer <3

[Alex Vaystikh @bornio]

@iruletheworldmo time to delete my subscription: friday

i am sick of it.

[Alex Vaystikh @bornio]

@DmytroKrasun So cool

[Alex Vaystikh @bornio]

@jul1anodev yes

[Alex Vaystikh @bornio]

I wanted a diff tool for rendered markdown similar to Linear’s but couldn’t find one.

So I built it: https://t.co/MT0BDyXbdT

[Alex Vaystikh @bornio]

@thenanyu This is gonna make T3 code so much better. Thanks, @theo !

[Alex Vaystikh @bornio]

@GergelyOrosz Waterfall and Agile were solutions to a problem that does not exist anymore: execution is expensive.

This means 2 things:
1. Ideas are expensive again
2. The trade-off between agile (ship small) and waterfall (ship big) is now zero.

[Alex Vaystikh @bornio]

Ideas are cheap. With AI, execution is cheaper.

Waterfall and Agile were solutions to a problem that does not exist anymore: execution is expensive.

Execution is now cheaper than ideas, and ideas are now expensive.

What this means for you is that testing your idea piecemeal is now bad advice. While you ship a small slice of your idea, I ship an entire product. While you collect feedback on a small feature, I collect feature on the whole idea.

In other words: the tradeoff between waterfall and agile is now zero.

[Gergely Orosz @GergelyOrosz]

Hear me out:

We hated "waterfall" because after Big Upfront Planning, implementation took FOREVER, and by the time we were done, the plan was out of date

But with AI, implementation is rapid (100x faster, or more, than before)

So Big Upfront Planning can now make sense?

[Alex Vaystikh @bornio]

both approaches are solutions to a problem that's largely solved now: execution.

Ideas were always considered cheap, relative to execution. Now, execution is cheaper.

Logically, ideas are now expensive, and entire ideas, not execution-capped slices, can be iterated on quickly.

Plan big!

[Alex Vaystikh @bornio]

@theo In T3 wen?