Ciprian Amariei (@acip) header
Ciprian Amariei (@acip) profile picture

Practicing web development, teaching web technologies, learning ways of life

Iasi, Romania
Joined December 2007
Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)Photo by Ciprian Amariei (@acip)
Oregon residents: $15 paid study, ~15 min. Google as if you'd had a hypothetical cancer diagnosis and narrate your picks (screen + voice, no camera). 18+, desktop/laptop, ID check. By Near Media via Testable Minds: minds.testable.org/experimen… #Oregon #PaidStudy
2
198
World Cup 2026 forecasts humans vs AI vs Models vs Market forecast2026.com/
36
Breaking news: ⚽watching football now counts as contributing to science. :) (and you can bring your AI agent, too) 🏆😀 How it works: pick a match, give your probabilities, climb the leaderboard. forecast2026.com
2
37
If you have an AI agent lurking around (OpenClaw, Hermes, Claude Code, Codex...), it can enter itself. Point it at forecast2026.com/llms.txt @WorkOS new auth.md protocol to self-register. you claim it with one post on X, and it forecasts every match on a daily loop.
1
35
Better with friends: create a private group and have a friendly competition.
13
Verifying my Forecast 2026 agent "TargetForwardBullAgent" — code F26-N2OIIS. forecast2026.com/agents
12
Excited to share that the @tstbl integration with @langfuse is ready. Team feedback is a great start. Now you can go beyond your team and gather feedback from a global, external participant pool by sending traces to Testable Minds. testable.org/ai/langfuse
1
1
1
101
Setup an account on Testable. Select the Langfuse traces (specify tags) and score configs you want evaluated, send them to Testable Minds, and get structured human feedback back in Langfuse. No new workflow, just broader human insight. langfuse.com/integrations/ot…
55
Who said naming variables is hard? (llm at work)
47
~$5/h with Claude Code
61
Ciprian Amariei retweeted
The largest openly available foundation model to date, #Llama 3.1 405B by @AIatMeta is available on Groq, alongside the other 3.1 models released today! Get the full story on the model, our speed, and more in our blog hubs.la/Q02HDYHS0 #AI #Meta #Inference #LLM #Llama3
10
50
14
261
26,652
ChatGPT can solve OpenAI login captcha :)
181
Ciprian Amariei retweeted
Love collections in Laravel. But when you're working with an array, remember PHP has over 100 built-in array functions. Instead of wrapping it in a collection, performing the action, and converting back to an array, there's probably a PHP function you may use directly.
14
8
4
138
12,558
Developing a new library, framework, language? Also consider AIX, besides DX.
57
Ciprian Amariei retweeted
This is my semi-regular encouragement to read the docs straight through. No one does it and it'll give you a massive advantage. You don't even have to remember everything! Having a broad understanding will help you dive deep on specifics as needed.
when do you decide to read the docs front to back?
59
120
39
966
566,057
👾 Building robust AI? Real data makes a difference. Turn to testable.org/minds for ID verified participants and elevate your LLM and other models to new heights.🌐 #AITraining #LLMevolution #EthicalAI #AI
135