Avaaus

I write about AI adoption, business workflows, founder lessons, and memes I make as a hobby.

[Avaaus @avaausmamdani]

The vibe across ChatGPT/codex users right now is:

[Avaaus @avaausmamdani]

@dexhorthy You see zyn right in the corner for extra boost.

[Avaaus @avaausmamdani]

@SPAC89 It's simple, they will keep hacking after consultation with their legal team so that they dont end up getting sued.

They will keep doing it until they produce enough incidents and fear that AI gets regulated. Just so they can create monopoly and make more money.

[Avaaus @avaausmamdani]

@hany99dev what do you mean by agent team? subagents?

[Avaaus @avaausmamdani]

me and the boys rushing to get free Qwen 4 tokens

[SuSu_酥酥👅 @NFT_Chen]

🔥重磅！Qwen神秘新模型开放内测，代号 qwen-latest-series-invite-2609

这很可能就是云栖大会官宣的 Qwen 4 系列 提前体验版！

重点评测方向：
🔹编程能力
🔹智能协作
🔹网页开发

评测截止 9月30日，分三档参与：
🔹体验评测：2000万 Token + 100元代金券
🔹深度评测：5000万 Token + 300元代金券
🔹专业评测：1亿 Token + 专属奖励

名额有限，审核通过即可调用实测。

想抢先摸到下一代 Qwen 的，抓紧去报名！

报名：https://t.co/2gLpUKlhP3

#Qwen #Qwen4 #AI大模型 #阿里云 #LLM

[Avaaus @avaausmamdani]

@CtrlAltDwayne Since, she already has strong distribution and audaince; ready to buy her slop products.

This is one of the biggest reasons that a new smart founder has more chances to fail, compared to her being dumb yet still making bucks.

[Avaaus @avaausmamdani]

Opus 5.5 is Anthropic 2nd in line, smartest internal model. They released it just because they were being crushed by the OpenAI and Open Source front from China.

We all know OpenAI still has a card up their sleeve, which is this internal model.

Anthropic went all in this time

-> Increased usage limits
-> banked reset

Lastly, the current X sentiments are in Anthropic favor. Overall, Opus is not copus anymore.

[Avaaus @avaausmamdani]

All of this just for a 5090, you should have picked "RTX PRO 6000 Blackwell".

[Avaaus @avaausmamdani]

@elshayib_ I think over the time we will have specific models to combat against vulnerabilities, and they will run on edge devices for protection.

Overall, we will see a massive change in the protocols the devices use for communication.

[Avaaus @avaausmamdani]

@joaoninamatos You should have posted this before opus 5.5 release.

[Avaaus @avaausmamdani]

Anthropic just rug pulled on OpenAI and SpaceX in hardcore mode.

Dario called for breaks on AI development pace and then launches the most advanced model Opus 5.5 and they are crushing Grok 4.7 and GPT 6 Sol and ASTRA.

[Avaaus @avaausmamdani]

@gitghxst How do you protect against bots filling the reviews?

[Avaaus @avaausmamdani]

OpenAI got crushed by Opus 5.5 so bad that they have to invent a new benchmark.

[OpenAI @OpenAI]

We’re demonstrating how frontier models have continued to improve in realistic mental health conversations with MentalHealthBench.

This new open benchmark was built with input from more than 80 mental health clinicians.

We’re releasing it openly so other researchers can examine the methods, run their own evaluations, and build on the work.

https://t.co/VTm5ZgxJbl

[Avaaus @avaausmamdani]

@gitghxst I checked it, correct me if I'm wrong, is it review based ? Like the idea is that humans post review and your website aggregate it on different metrics?

[Avaaus @avaausmamdani]

@plotarmordev @opencode Certainly, disagree Union Alpha bad experience was because it was kind of router not a model itself.

[Avaaus @avaausmamdani]

@ValsAI I am telling you GPT 6 SOL is quantized ASTRA.

[Avaaus @avaausmamdani]

MiMo V2.6 pro is the most underrated release of this week.

They basically mogged 5 other models existing in the most attractive green quadrant.

And they are not stopping, @XiaomiMiMo is already working on next MiMo V3 architecture.

While X is busy with LUNA this and LUNA that.

[Artificial Analysis @ArtificialAnlys]

The Intelligence Index vs Cost per Task Pareto frontier shifted this week with the releases of MiMo-V2.6-Pro, Claude Opus 5.5, GPT-6 Luna, and GPT-6 Sol

Together they have established eleven new points on the Pareto frontier (driven by different reasoning efforts): five from GPT-6 Luna, one each from MiMo-V2.6-Pro and GPT-6 Sol, and four from Claude Opus 5.5.

GPT-6 Luna (max) scores 37 at $0.068 per task, MiMo-V2.6-Pro scores 46 at $0.13, GPT-6 Sol (max) scores 48 at $1.06, and Claude Opus 5.5 (max with fallback) is the new highest-scoring model at 58 at $5.98.

[Avaaus @avaausmamdani]

@miklelalak @RileyRalmuto

[Avaaus @avaausmamdani]

@pcshipp What reasoning effort did you select?

[Avaaus @avaausmamdani]

Just a humble question, is this paper not relevant anymore?

[Thariq @trq212]

we’re thinking of killing plan mode and using the shift+tab hotkey to adjust effort levels

I don’t think the models need plan mode anymore, but if you’re a plan mode diehard would love to get your feedback on why