Xnd

@elliotarledge im satisfied using glm, it's so good on agentic coding

glm 5.1 feels like a healthy mix of claude and codex

@kitlangton @opencode please why i can't copy pasting on opencode TUI? I can't block text

I know many of you have been long awaiting this monumental upcoming @opencode feature. I'm so happy to announce that: Yes, when you go bonk on the logo it will respond with delicacy and suppleness to the current theme.

@thdxr why not you add more $15 $20 plans man,everyone have their own budget

we've had a huge spike in OpenCode Go subscribers so we have people on the team entirely focused on securing more capacity we're growing faster than our providers are receiving GPUs we have options but might be a bit bumpy as we figure it out

@dhanyindraswara asli,pindah ke github copilot pro + opencode go,jos coy

$20 Anthropic subscription.

@jayair @kitlangton pls,i like opencode's TUI,but why it's so hard just for copy and paste from this TUI make it easy please,even i can't paste long prompt using shift + insert.

Kit built this, should we ship it? (sound on)

@netanyahu low morale creature shit

Israel under my leadership will continue to fight Iran’s terror regime and its proxies, unlike Erdogan who accommodates them and massacred his own Kurdish citizens.

@jayair @kitlangton hell yeah

@setyamickala bingung mau garap yg mana jir,saking banyak dan saking ga yakinnya

Kalian masih garap airdrop ?

@skyl3r77 this browser is fucking shit man,i use defi from this browser,and my metamask can't connect because there's brave wallet on it,i prefer vivaldi, it's best for speed

when will people realize this browser is shit

@yappingfess lu udh mandi dan ga kena najis aja itu sebenarnya udh gaperlu wudhu lagi,udah suci

yap! islam only sebenernya kita tuh boleh ga sih wudhu dalam keadaan bugil sehabis mandi?

@7a7zz can you make a post about how to use it? I'm also a budget seeker plan,confuse what to buy,and I see this article,maybe I will consider your GitHub co + opencode recommendation,I hope you make a post how to use it,or if it's simple,please tell me about it on reply

@iniheksa thanks bang sumber belajarnya

Bayangin, lu ngk ngerti tokenomics selain diagram alokasi airdrop, tapi langsung apply sebagai Tokenomics Analyst. Lalu di hari pertama kerja, lu langsung disuruh review tokenomics Bittensor (TAO), Starknet (STRK), dan GMX. Isinya? Banyak rumus matematika yang bahkan nggak bakal lu temuin di mana pun selain di whitepaper mereka sendiri. 🤯 SS: Respons CEO waktu gue share deliverable pertama. 😂 Tapi setelah seminggu, gue justru suka kerja bareng dia. ➤ Super suportif: ➤ Kursus dibiayai ➤ AI subscription dibayarin ➤ Dikasih ruang buat belajar dan berkembang ➤ dll Buat lu yang tertarik belajar tokenomics dan belajar reseach, bisa cek website berikut: 1. onchain.org/magazine/crypt… 2. tokenomics-learning.com/en/home/ 3. udemy.com/share/107Pdk/

@ibocodes I need recommendation,codex,Alibaba or glm?

chinese models used to be the cheap alternative why are they priced like claude & codex now 😭

@tomhacks Yeah,Im using Gemini ai pro,I got discount and I try it,man, it's worst than chat gpt,at some point it's good,but hell no,they said Gemini 3.1 pro is good,but I don't find it where this model is good,or just me that dontknow how to use it? I already use all Gemini coding platfor

> Buy Gemini AI Ultra plan > Excited about the increased quota > Testing Gemini CLI - first message Error 429 No capacity available for model gemini-3.1-pro-preview on the server > Feels-bad - where is gemini-3.1-pro-high that I actually want to use? Antigravity > Installs IDE and asks agent to explain harness > Finds: STOP and wait for the user's explicit approval before proceeding to execution. Smallest possible steps everywhere... does not run autonomously, no working ralph loop, no headless mode. > Finds Agent Manager, tries is out - same limitations - no automation, no headless, no orchestration --- Who thought this would be a good product to release? How is that supposed to compete with frontier coding harnesses if the CLI does not have the newest model, the servers don't want to serve it and the IDEs minimize usage as much as possible. $124.99 (from $249.99) worst investment I made - cancelled

@cipherwebthree gimana cara make ya?built in langsung kah

CLAUDE KELUARIN FITUR BARU BIAR LU GA GAMPANG KENA LIMIT!! Pasti kalian semua keluhannya kalau beli Claude itu limit kan wkwkwkkw, termasuk gua juga ngeluhin begitu. Nah, kali ini ada fitur baru yang dirilis sama Claude, dimana kalian bisa minta task dilakuin sama Sonnet / Haiku, tapi nantinya tetep akan ada peran Opus sebagai Advisor/Penasehat. Buat yang mungkin belum tahu, model Claude itu ada 3 : 1. Haiku Haiku ini bisa digunain untuk tugas sangat ringan, model Claude yang paling bawah kelasnya 2. Sonnet Nah kalau Sonnet ini cocok untuk tugas sehari-hari, dan jadi model terbaik di Claude untuk kalian yang Claude nya free 3. Opus Kalau mau pake Opus, lu harus join Claude Pro dulu, sebulan yang paling murah itu $20. Nah kalau Opus ini model yang terbaik, tapi ini yang bikin token lu cepet abis. Dengan fitur baru ini, Sonnet yang ngejalanin tugas lu, tapi dia minta Opus buat jadi penasihatnya. Ini diklaim outputnya 2% lebih bagus dari output Sonnet biasa, tapi bisa hemat token lu berkali-kali lipat daripada lu pake Opus. Ini masih beta dan udah live di Claude, bisa kalian cobain dan semoga gak kena limit terus ya WKKWKW

@ullquiorrac sok tau opinion

Jangan terlalu ngarep HP android bisa dipake jangka panjang (4-6 tahun) termasuk yang flagship sekalipun

@Elektroo12_ parah,orang dewasa aja rata rata 20x udah mulai Ngos ngosan😢

Pasti dia kesakitan😭

@keluhkesacrypto duitnya gede

Katanya Windah Basudara bisa dapet 2 SOL setiap 2 hari yang berarti 1 SOL/Hari, dapet darimana tuh? staking kah?

@ceogalxe speknya harus make ram 32gb kah? atau 16gb aja cukup bang?

COMBO MENARIK = GEMMA 4 + OLLAMA + OPENCLAW (FULL LOKAL) ‼️ jadi ini setup bikin AI agent pribadi yang : 🔹 jalan di laptop kalian (local) 🔹 bisa coding, automation, dll 🔹bisa diakses lewat Telegram 🔹 100% FREE , NO API, 0 data keluar 1⃣STEP 1 🔹Download & Install : ollama.com 🔹Install kayak biasa (next-next aja) 🔹Support : Windows / Mac / Linux Cek install : Buka terminal / CMD, ketik : ollama --version kalau keluar version, aman. 2⃣STEP 2 — Download Model Gemma 4 Di terminal, jalanin : ollama pull gemma4 Kalau bisa chat = DONE 3⃣Install OpenClaw, pilih Ollama jadi provider 4⃣Sambungin ke Telegram! jadi kalian bisa akses agent dari mana aja, dan tetep run di local ! Cuma pake satu prompt, langsung jadi SEO calculator yang jalan HTML, CSS, JavaScript, tanpa ribet debug. Benchmark coding naik jauh, dari 110 jadi 2.150 dibanding Gemma sebelumnya.

@bukanpamanmu bang mending codex apa cursor ya yg plan $20 an?

🧑‍💻 Sebagai Full Stack Software Engineer (sejak 2014), Paman kini menyelesaikan berbagai bug di GitHub jauh lebih cepat berkat berbagai jenis AI/LLM yang digunakan Dulu butuh berhari-hari. Kini hanya hitungan jam. AI/LLM telah menjadi partner kerja yang luar biasa, meningkatkan produktivitas gw secara signifikan. Bagaimana pengalaman Kamu menggunakan AI dalam pekerjaan teknis? Apakah Kamu juga merasakan percepatan yang sama?