- Tweets 829
- Following 304
- Followers 726
- Likes 120,783
Claude just spent 3 hours recovering all of this and now I have this cluster fuck on my desktop to deal with.
Mac iCloud file sync is so fucking annoying. It always messes up git processes. I finally decided to turn it off completely and it disconnected all the repos and moved everything to an archived folder. like why the fuck would that even be a feature.
2
203
Mac iCloud file sync is so fucking annoying. It always messes up git processes. I finally decided to turn it off completely and it disconnected all the repos and moved everything to an archived folder. like why the fuck would that even be a feature.
2
295
I asked Claude Opus 5.5 to build a typing game end to end, from one prompt. (prompt below)
outpace.ethanplus.ai/
2
4
198
Prompt: Build a fun, extremely smooth typing game end to end. It should measure your speed but be genuinely fun to practice with, and it should adapt up and down to your skill so progression feels fair. Free website, global leaderboard, optional GitHub or X username to show your avatar (no sign-in). Fully responsive. Visually test everything, test the mechanics, and keep iterating until it's something you're proud to show off as a demo of Opus 5.5. Use Sonnet 5.5 subagents for well-defined parallel tasks, track usage and timing separately, and open it in my browser when done.
1
69
Ethan+ retweeted
estamos vivendo um período em que a pessoa inteligente precisa permanecer em silêncio para não ofender os ignorantes
325
27,376
123,905
1,365,268
Ethan+ retweeted
This is actually a genius way to weed out liabilities before they become employees who will most certainly sue you for alleged grievances.
I had an offer extended to me from JPMorgan like 4-5 years ago and I was on the phone with the hiring manager who had a thick Chinese accent and the first thing he said over the phone was “Sorry the process took so long. Our HR lady is SO RETARDED” and I was like oh ummm
38
233
21,868
3,608,475
Ethan+ retweeted
If you're using AI all day and you're not making more money, it means you have the same problem you had before the AI: you're doing stuff that doesn't matter.
This is the real reason 99% of people don't get ahead. They don't know how to tell the difference between what's important and what's irrelevant. And just doing irrelevant things faster gets you the same outcome: nothing.
Except this time it cost you tokens.
406
715
9,621
374,025
wtf. Claude deleted all of my 24 resend api keys as a part of a routine security audit two days ago resulting in every production application silently failing all emails...
It didn't have access to Resend directly so it somehow used my normal browser?
4
4
218
Just got the first customer order. Absolutely insane that I have not touched any part of this. I have only prompted and set up payment accounts.
I had GPT-6 build an entire canvas brand and gave it $2k/mo to spend on ads until it’s profitable.
inkrally.com
Ads are just starting to warm up so we’ll see what happens.
2
2
160
600 stars in 3 days. nice
github.com/ethanplusai/astra…
1
3
147
This is where it all started…
The new OpenAI tool is insane. It can produce code or explain anything based on user prompts. People who start using this will have an extreme competitive advantage over those that don't.
chat.openai.com/
117
Built a fully automated outreach system that costs <$25/mo to run outreach at scale.
- builds prospect lists with verified contacts
- agents handle custom responses instead of static workflows
- emails, SMS, LinkedIn, and calls
- runs local
Will opensource on Github asap
3
313
I reduced my GPT-6 Astra usage by ~94%.
Astra plans. DeepSeek Flash builds. Astra reviews.
Same quality of work, just less codex usage.
1
4
981
Per 1M tokens:
Uncached input: Astra $10 / Flash $0.15-0.30 (33-67x)
Cached input: Astra $1 / Flash $0.003-0.006
(167-333x)
Output: Astra $50 / Flash $0.60-1.20 (42-83x)
92-93% lower estimated compute cost overall.
1
21