oitozero (@oitozero) header
oitozero (@oitozero) profile picture

Father of 2. Former tap dancer. From C to code whisperer. Recovering CTO, aspiring founder. 3:48 to 2:53 marathoner.

Home
Joined April 2013
Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)Photo by oitozero (@oitozero)
Fable 5.2 charged me $0.73 to work out 2+2. Opus 5.5 charged $0.38. Fable 5.2 came out this week and I ran a small test. Same prompt to both models, a single arithmetic operation, nothing else. The Fable cost was almost double than Opus.
2
1
54
WTH? Claude Meetings? Anthropic, I know the IPO is coming, but you need to slow down.
Introducing Claude Meetings: add Claude to your team’s video call and it builds docs, whiteboards, and other artifacts with you while you’re still talking. Starting to roll out in beta on Team and Enterprise plans.
34
I expect plenty of teams to bump the default to the newest model and send every request through it. The answers still come back right, so nobody notices the waste until limits are hit or the invoice arrives.
1
10
Before reaching for the newest model, ask what the task actually needs and pick the cheapest one that does it well.
4
I made a Claude Code mod that shows your PRs that are ready to merge, right above the prompt. /plugin install ready-prs --marketplace oitozero/ready-prs github.com/oitozero/ready-pr… #Claude #Claude_Code
2
45
Three settings make it work in busy team repos: • scope: mine, so only PRs you opened show up • readiness: clean / clean-or-unstable / approved • reviews: on or off
12
Not cool. Claude Code keeps changing my effort to high in new sessions. Already tried /effort and .claude/settings.json. @ClaudeDevs
1
1
88
"Out of usage credits. Buy credits to keep working" Apparently, Claude Code is on an unpaid break until further notice. #ClaudeCode
2
86
Great companies almost always have great names. The losers in the same space usually don’t. Facebook 🚀 vs Myspace ❌ Netflix 🚀 vs Qwikster ❌ Tiktok 🚀 vs Quibi ❌ Spotify 🚀 vs Rdio ❌ Coincidence, or does the name actually decide more than founders admit?
32
"It can take up to 48 hours to be reviewed. You'll get an email when the review is complete."
36
Best use of Opus 5.5 yet. Jev might be making the Valencia call for me.
38
If you want to understand how Jev works, read this.
I've had many people reach out saying they've read everything they can about Jev but still don't get it. Here's an explanation for the normies
47
oitozero retweeted
we just finished a 90 day GEO experiment that most people in AI search will hate we took a partner from 4% to 80% visibility, pushed BlackRock out of the answers (with decades of authority and $15T under management) and didn't publish a SINGE piece of CONTENT the entire time no blogs, no landing pages, no "AI optimized" articles, literally nothing now our partner is showing up ahead of multi-billion dollar legacy brands like BlackRock, KKR and Temasek in LLMs I want to explain why we ran it this way, because I think most people in GEO are spending their money on the wrong thing if you look at every tool in this space, the product is almost always *content* if you look at every agency pitch, it's some version of "we'll write 40 articles a month for you" and the reason is quite simple content is simply the easiest thing to produce, sell and the easiest thing to invoice for but what nobody had actually tested and questioned is whether it's the thing that moves the needle or whether it just happens to be the thing everyone is doing! so we picked a brand new partner, and didn’t deploy a SINGLE piece of content for 90 days and pulled on every other lever we had (we really wanted a clean answer) now, here's the thing you need to understand about how these models work (and think) instead of reading your blog and deciding you're the best in your category the LLM builds a picture of you from everything it can find and most of the weight goes to what other sources say about you, not what you say about yourself your own content is literally the least trusted input, because you're the one claiming you're great and the model knows that! so instead of writing, we spent the 90 days on the stuff that actually feeds that picture 1) the first thing we fixed was the entity itself the partner was described in six different ways across the web depending on where you looked, and models punish that kind of inconsistency hard. we went through and made every source agree on who they are, what they do and who they serve. 2) then we mapped the sources for every buyer prompt in the category, we pulled which domains the models were actually citing, and then made sure the partner was present and correctly described on all of them. directories, databases, industry lists, wikis, news. if a model pulls from a source and you're not on it, you essentially don't exist for that question 3) after that we worked on third party validation, so mentions and references on sites the models already trust, framed around the actual questions people ask we also cleaned up the structured side, schema, naming, entity data anywhere a machine reads it. boring but it matters! 4) and the whole time we tracked the real buyer prompts, reverse engineered the sources behind each answer, and went after it one by one the combination of all these is what took them from 4% to 80%. companies with billions in brand spend were simply pushed out of the answers by someone who didn't add a single new page to their site! think about that next time someone's convincing you that it's all about content also, the lesson isn't that content is useless content works, but only once the model already believes you're credible! most people are doing it backwards and building the roof before the foundation exists, and then wondering why 40 articles a month isn't doing anything meaningful if you want the exact playbook we used on this, RT + comment "Visibility" and I'll send it over
40
7
9
40
10,924
why is everyone asking to drop their startup url?
34
99 days straight of chess puzzles. i expected +10 IQ. instead i got measurably better at solving chess puzzles and absolutely nowhere else in my life.
26
This @chamath video will be a master class in life/ buisness on what failure is or isn’t. Because There is No Failure. Must hear for everyone:
40
294
12
1,716
185,932
My Uber driver asked what I do for work. Software.... Cool. Can you look at something? He handed me his phone at a red light. Terminal. Claude chat. Green P&L. +$6,200. He drives Uber 4 days a week. Makes $1,100. Has a 2-year-old daughter. Where did you find this? Your article. The 10,000 wallets one. He read it 3 months ago. Did not understand half of it. Asked Claude to explain it like he's five. 214 messages. All during breaks between rides. Parked at gas stations. Waiting for pings. 1st thing Claude told him: 87% of wallets lose money. Do not be the 87%. He installed poly_data. Fed it to Claude. Found 47 wallets with Sharpe above 2.0. Filtered crypto only. Quarter Kelly. $200 starting bankroll. From his tips. 93 messages later Claude helped him build the 20-line brain from the article. Bayesian updates. EV filter at 5%. Fully automated. Last 45 days: → 480 trades → 91.3% win rate → +$6,200 Best trade, whale convergence on Fed rate cut. 4 wallets entered in 2 minutes. Entry $0.12. Resolved $1.00. +$1,760. While dropping off a passenger at JFK. The passenger tipped him $5. The bot made $1,760. His wife found the Telegram alerts on his phone. Thought he was texting another woman. He showed her the P&L curve. Can you make me one? How long until you quit driving? He looked at me through the rearview mirror. I'm not stopping. Uber is my cover story. I wrote the article. He actually opened terminal. You only need Claude + laptop + 1 hour/day. Giving This Free for 24 hours. To get it: 1. Comment the word "Trade" 2. Like and Retweet this post 3. Follow me @codewithimanshu (so i can DM you) Save this post. Deploy the bot this weekend. Start with $200. Scale on evidence.
235
150
2
332
45,095
oitozero retweeted
Them: Since the good old days of the mid 20th century, a few people have gotten richer and everyone else has gotten poorer. The actual data: xitter.gq/i/status/1665194196469…
Replying to @Duderichy
This is my favorite animation!!! I have it bookmarked on my phone for parties. Such a whitepill on past 40 years.
181
229
34
2,131
1,335,811
This is Baremetrics revenue over 9 years, which doubled since their acquisition by a PE firm. This should be framed and titled: “SaaS outrage doesn't matter”
11
1
70