Interconnects.ai coverage reviews rapid open AI model launches including Gemma 4, DeepSeek V4, Kimi K2.6, MiMo 2.5, and GLM-5.1 against CAISI V4 assessments

new artifacts!

i also comment on the open<>closed model gap, where US CAISI and @EpochAIResearch disagree, arguing that both are incomplete: for an assessment of the very frontier, we must elicit the best performance by tuning prompts and harnesses with the models

@adam_fresko grok will take a little longer sadly.

@elvisofdallas @iruletheworldmo Codex being that the mobile app hasn’t rolled out for windows. I also used it for Claude desktop before remote control came out.

I do stock trading so I like the think or swim desktop app over the mobile

@DanielWhit21874 @iruletheworldmo ??? just use open design ... https://github.com/nexu-io/open-design

@iruletheworldmo Calling it now: 5.6 will cost 1.5x more tokens than 5.5, so 3x usage per token over 5.4. $100 a month plan is the new $20 a month plan

@xeophon @EpochAIResearch open vs closed is a trap.

@DanielWhit21874 @iruletheworldmo This is the dashboard it made, GPT5.5 under the hood. I said when coming up with this idea, I said "I don't like ChatGPT slop, make sure it's nice and tight App style UI". That was it...

@iruletheworldmo Seriously, how do you guys know this stuff?

@iruletheworldmo Nah, its a routine now, I'm still waiting for that one small chinese disruptor model though.

There's just so much. I don't know where to begin. You notice day 2-3. If not you'll definitely see the difference by day 7. Persistent memory, better tooling, 50+ Skills built in, it seemed to know stuff about me when I didn't even give it info, it gleaned, made correlations, and was smart enough to get the bigger picture. It talks in my voice. I didn't swear at it, I generally don't swear around agents unless it's chatgpt... but by day 3 it must have figured out what posts were mine and started swearing, being blunt. Talking the same as I do. it's refreshing after listening to ChatGPT brown nosing BS.

@iruletheworldmo I really want gemini 3.5 to be good 🤞 if it is, it will be a game changer. Do you think it will be a huge enabler for agentic harnesses?

@iruletheworldmo I’m curious whether gpt 5.6 actually will be a better model for front end

@striver_777 this week.

@iruletheworldmo Chrome Remote Desktop from my phone has controlled 2 of my home pc’s for months

@MichaelKraw2005 yes. elon is a beast.

@adam_fresko @iruletheworldmo Elon said they've finished training for the 1.5 trillion v9 model and 3-4 weeks for cursor data and rhlf and full release etc

@iruletheworldmo It is Sunday now. Do you mean actual next week or this week?

Love it - thank you for taking the time to share with me (or all of us!)

I loved Dispatch when it came out - I’ve been thinking of whipping up a web app that can queue prompts to multiple AIs on my desktop so I can do more to keep processing moving along

Do you use a keyboard with your phone for remote control?

@adam_fresko @iruletheworldmo He said one month but I guess by his usual timing it probably will be the start of July rather than end of June

@iruletheworldmo Do the insiders think there’s any chance for XAI