CheckMate

@fabiohak21 @nahcrof @nahcrof I am ready to buy in, but please add the mimo models.

@nahcrof hey, any news about Mimo models? please

I'm getting close to running a database entirely in-memory

@paulg Doubt any western sources and always cross check. just this week the Russians took over 11 populated areas.

What changed three months ago?

@simonbrown Thanks, love it! I put a high premium on shorter, condensed books that are respectful of my time! Ideally 100-120 pages. See if you can cut even more!

@sladurb I haven't seen the templated manuscript yet and the figures still need to be redrawn, but perhaps somewhere in the region of 130-150 pages. A short book, but significantly longer than the self-published Leanpub version.

Not long now ... the first round of copy editing is complete! 🎉

@TechLayoffLover Facts- free posts! These unfortunate characters do not exist.

$247,832 in debt and counting A computer science grad from Carnegie Mellon with a 3.87 GPA Class of 2025, recruited by Google, Microsoft, and Meta his sophomore year All three offers rescinded "due to economic headwinds" 1,847 applications sent since graduation 47 interviews 12 final rounds Zero offers His roommate from Stanford drives DoorDash 14 hours a day making $11.23 an hour after gas Another classmate from MIT stocks shelves at Target for $16 an hour The career services office still runs LinkedIn ads about their "94% placement rate" Based on 2019 data Before AI murdered the entire profession He spent four years learning React while React developers were being replaced by Claude Learned Python while Python jobs got automated by Copilot Learned algorithms while algorithms were writing better algorithms than humans The ultimate fucking irony He graduated summa cum laude in the exact skill that just became obsolete Carnegie Mellon's computer science program costs $61,958 per year For a degree that's now worth less than a community college welding certificate But sure, keep telling kids to learn to code

@opencode it affects all other resellers, the issue is with the provider you use

we've had intermittent issues with kimi 2.6 on OpenCode Go we're investigating with our providers, sorry about that hopefully can use one of the other models in the meantime

.. which is why @nahcrof does not offer Minimax 2.7, I guess ? I am still rooting for such providers, hope there is another way.

MiniMax shoes why - soon all oss Chinese models will include license clauses to prevent unauthorized resellers. The have to prevent prices rot for their models, now that they all took the VC poison pill.

@nahcrof is very cool - it is a dirt cheap , quantized cluster of oss AI models run by an honest founder on leased hardware. The bet is a "race to the bottom" in token pricing. Unfortunately, it is a "house of cards" business ...

@nahcrof The opencode section in the docs does not list any deepseek v4?

Alright, CrofAI is officially the cheapest option for deepseek-v4-pro have fun saving tons :)

@jun_song Yes, that is how it works, unfortunately. MiniMax is showing the way. Coincidentally, most Chinese got VC finding.. never a good sign!

현재 저렴한 중국의 AI도 언젠가는 가격이 오를겁니다. 프론티어의 모델과의 성능 격차때문에 저가 정책을 유지하는것 뿐입니다. 성능을 따라잡는다면, 그들도 가격을 올릴수밖에 없습니다. Seedance 2.0이 그의 완벽한 예시입니다. SoTA 레벨을 달성한뒤에, 가격은 터무니없이 비싸졌습니다.

@abukinansafadi3 @jun_song News link?

@jun_song You are not following the news. A flood of huawei gpu will soon hit the market

@zeddotdev 3-way git merge review .. was the reason i uninstalled zed 2 months ago

What's that ONE feature you're missing in Zed?

@no_stp_on_snek Looks awesome! Zero python please!

vllm-swift v0.2.2 is out hybrid models (qwen3.5, qwen3.6, qwen3next) now scale with batch instead of going flat at B>1. ~16x total tok/s at B=64 on qwen3.6-27b on M5 max. also fixed in this release: - OOM at 4B / B=64 / 8K context - crash at prompt length ≥ 2048 with batched decode - source install no longer needs manual mlx.metallib copy prebuilt brew bottle so no swift toolchain needed: brew tap TheTom/tap && brew install vllm-swift readme + getting started: github.com/TheTom/vllm-sw…

@opencode @thdxr - $20 Go Pro when? Do not make me go for Kilo Pass or OpenRouter, please?

@jun_song Keep posting these rankings regularly, please!

4/25 기준 카테고리별 최고의 AI: 성능 : GPT-5.5-Medium 가격 : Deepseek-V4 가성비(코딩) : Kimi-K2.6 가성비(에이전트) : Minimax-M2.7 최고의 로컬AI : Qwen3.6-27b 이미지 : GPT-Image-2 영상 : Seedance 2.0 로컬 이미지 : Ernie-Image-Turbo 리서치 : Grok-4.3

@minilek @ylecun Your analysis is both correct and wrong: - ai will impact CS job demand, but over a longer timeline - the collapse is part of broader economic decline, mostly the end or 0 interest rates and ... We have not had a recession since 2008.

Many folks seem to be confused, and think the collapse of the CS major graduation numbers at Berkeley could be linked to the "AI is taking SWE jobs" hysteria narrative. Here's the easiest way to see that this is false: the timeline doesn't fit. The graduating class in 2027 (first small CS cohort graduating) has students who arrived on campus as freshmen in Fall 2023, with freshman admission targets set (i.e. shrunk) by the university in Fall 2022. So, the hysteria narrative obviously doesn't match the timeline; ChatGPT didn't even come out until November 2022. Now consider the plot below; orange curve is what % of bachelor's degrees are CS degrees each year at Berkeley, and blue curve is what % of applicants applied to be a part of that graduating year, intending to be a CS major in their application (combining both junior transfer and freshman applicants). In other words: * orange measures CS graduate production * blue measures CS demand (via % of all applications to the university) What do you notice? The collapse in orange (CS grads) isn't because of a collapse in blue (demand). In fact it's the opposite: orange collapsed at a time when blue was going up. 1/

@thomasgauvin Wait, I can't send email with just smtp protocol, only your privately rest api?

The wait is over. Cloudflare Email Service is now in public beta 📧 Send and receive emails directly from Workers or REST API with global delivery on Cloudflare's network And just in time for you to build email agents with the Agents SDK!

npm uninstall -g @qwen-code/qwen-code

@Alibaba_Qwen killed the free/generous coding plan in qwen cli as of April 15th, 2026. No announcement! @MiniMax_AI 2.7 no longer free oss. Gemini CLI pulled the Pro models this month too. Kimi 2.5 and GLM 5.1 - the last 2 still free, brutal VC capitalism from now on!