Andrés Rodríguez (@mixedCase) profile picture

I like making computers do stuff.

Uruguay
Joined October 2014
Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)Photo by Andrés Rodríguez (@mixedCase)
Andrés Rodríguez retweeted
Replying to @4A4133
So instead 2048-bit prime they used one.fucking.byte as "secret"? And that bug has been out there since 2018 & nobody even bothered to do checks, tests, formal verifications etc? Seriously, fuck off you all. Never talk to me again about ai slop when horrible human slop shit like this exist out there. Talk to the hand.
8
19
1
361
13,551
Andrés Rodríguez retweeted
AI is a mirror. lazy people get lazy results. curious people get superpowers.
150
910
95
8,604
142,921
Andrés Rodríguez retweeted
Someone 'accused' me of vibe coding the LFRR simulator. I've been coding since 1982, I've hand assembled machine code, injected fixes into running processes, worked on software that's made it to Mars and worked as a software developer for decades. OF COURSE I USED AI CODE TOOLS
381
731
75
20,245
796,356
This is roughly 20% of my feedback cycles to the LLM at this point. As far as LLM verbal crutches go, this one's alright.
2
191
Andrés Rodríguez retweeted
18
289
15
5,682
119,832
42
98
21
1,231
231,911
Andrés Rodríguez retweeted
I told Claude to one-shot an integration test against a detailed spec I provided. It went silet for about 30 minutes. I asked how it was going twice and it reassured me it was doing work. Then I asked why it was taking so long:
234
241
169
7,693
1,225,541
I grieve for the loss of the em dash. I could not have started incorporating it into texts I actually put effort in at a worse time.
1
348
Andrés Rodríguez retweeted
Finally it's out, my experiment in all-monospace web! owickstrom.github.io/the-mon… I've tried to include a bunch of different elements to showcase the idea. Many more things could be added and supported. I've had a lot of fun with this so far!
149
337
80
3,237
205,259
Andrés Rodríguez retweeted
Still one of my favorite robotics use cases: Peanut Robotics cleaning hotels and workplaces autonomously. Need a home version of this guy
65
195
66
1,704
236,120
Andrés Rodríguez retweeted
130
11,986
402
185,437
5,587,625
Andrés Rodríguez retweeted
Google Calendar ask why a meeting was scheduled
70
477
90
7,591
716,136
Andrés Rodríguez retweeted
-----BEGIN PRIVATE KEY-----
AAAAAAAAAAAAAHHHHHRHRGRGRGRRRGURBHJB EORWPSOJWPJORGWOIRGWSGODEWPGOHEPW09GJEDPOKSD!!!!!!!!!!!!!!!0924QU8T63095JRGHWPE09UJ0PWHRGW
41
485
36
3,788
232,191
Andrés Rodríguez retweeted
OH: "This microservice could have been a SQL query"
62
742
105
6,227
Was proud enough to be able to pull off a decent shot of espresso on my new set-up that I felt like tweeting today, for a change:
4
Andrés Rodríguez retweeted
Compositionality allows a small change in requirements to translate into a small change in code. Compositionality is the single most important factor in determining the cost of code maintenance.
3
17
88
Andrés Rodríguez retweeted
Business logic is the layer of your application that no other layers depend on. It's the center of the onion. In testing business logic, you need only mock / fake / stub its dependencies, never the business logic itself. Business logic has no downstream dependencies.
6
8
1
49
Andrés Rodríguez retweeted
Consulting service: Your 25-person company is drowning in complexity with microservice architectures, serverless, and SSR. I say "your app will fit on two Heroku Standard 1x dynos and one Standard 0 database. $150/mo total. Eat the 100 ms of extra latency." You pay me $10,000.
41
220
33
1,721
Andrés Rodríguez retweeted
Why your big co will lose to a startup: 1. Organizational debt. You can't gather, learn from & execute on market data quickly. 2. Tech debt. You ignored countless engineers, now your tech asset has become a liability. 3. Personnel. You lost all the people who could fix 1 & 2.
13
66
12
356