pyk retweetledi
pyk
13.1K posts
this graph perfectly shows how hyped i was for @FireworksAI_HQ fire pass v2
so far so goooood, <1s ttft & >150 tps. hit a few rate limits on the first day, needs to adjust my workflow and my harness retries need some work
aside from that, zero issues. barely touching GPT 5.5 now
English
seumur umur baru 2x subs claude pro, setelah itu gamau subs lagi karena mahal dan cepet limit, akhirnya ngulik opencode dan bikin setup agent kaya gini pake model china & ChatGPT, cukup powerfull buat kebutuhan coding backend web3, research, refactor, brainstorm, etc.
sekarang lagi sering pake model deepseek v4 pro karena lagi murah tapi bagus bgt buat coding backend + web3.
Indonesia
Be honest — how many local models do you have downloaded right now? 👀
English
a few updates, more to come tonight (in ~9hours)
- sign ups are back! (early 2000s lookin' captcha but it works)
- deepseek models should have better fallbacks (deepseek-v4-flash seems to be struggling but it was meant to be in beta anyway)
- /usage-api/ returns "requests_plan" (the total daily limit of your plan)
English
rumors are circulating that DeepSeek and Kimi are merging.. the two most dominant open-source AI labs in China..
if this is real.. China just did what it always does when an industry becomes strategic.. in 2015 they merged CNR and CSR into the world's largest train maker overnight.. did the same with Baosteel and Wuhan Steel.. same with their nuclear and telecom companies
they don't let their best labs compete with each other.. they combine them into one thing and aim it at everyone else.
meanwhile OpenAI is suing Elon.. Elon is suing OpenAI.. Google and Anthropic are poaching each other's teams.. billions being burned just fighting internally before anyone even looks east
the sanctions were supposed to slow China down.. all they did was give them a reason to stop splitting the talent and start sharing it
DeepSeek@deepseek_ai
🔥DeepSeek Input Cache Price Drop! Effective immediately, the price for input cache hits across the ENTIRE DeepSeek API series is reduced to just 1/10th of the original price! Build more efficiently for less. 📌Reminder: The DeepSeek-V4-Pro 75% OFF promotion is still active until May 5th, 2026, 15:59 (UTC Time).
English
Ghostty is leaving GitHub. I'm GitHub user 1299, joined Feb 2008. I've visited GitHub almost every single day for over 18 years. It's never been a question for me where I'd put my projects: always GitHub. I'm super sad to say this, but its time to go. mitchellh.com/writing/ghostt…
English
@DeepInfra Why is glm5.1 more expensive than glm5, both models are same size and architecture
English
Price drop alert on DeepInfra:
• GLM-5.1 → $1.05 / $3.50 / $0.205
• GLM-5 → $0.60 / $2.08 / $0.12
• MiniMax-M2.5 → $0.15 / $1.15 / $0.03
• Gemma-4-26B-A4B-it → $0.07 / $0.34
(in / out / cached, per 1M)
deepinfra.com/models
English