Mino

AI news, research and bad takes. accelerationist.

professional shitposter in training

per aspera ad astra

[Mino @compounded_mind]

Kimi K3.1 might be coming soon

DeepSeek just updated V4.1 Flash and it beats K3 on 10 out of 11 benchmarks in DeepSeek’s own comparison.

would make sense for Moonshot to get its next model out as soon as possible

[SuSu_酥酥👅 @NFT_Chen]

🚨重磅预警！Moonshot 前端 JS 提前泄露 Kimi K3.1！K3.1 或即将上线！

官网模型选择代码里已经出现 k3d1-agent（Kimi K3.1）+ OK Computer 的 make3d 模式。

重点如下：
🔹K3.1 大概率已进灰度或即将官宣
🔹强化 3D 生成能力（游戏/交互场景直接起飞）
🔹长程 Agent 再升级，国产旗舰要加码了

Kimi K3.1也开始动手了，GLM、Qwen 你们呢？！

#Kimi #KimiK31 #国产AI #AIAgent #3D生成 #Moonshot #杨植麟

[Mino @compounded_mind]

@LexnLin fucking amazing

[Mino @compounded_mind]

@chetaslua happy birthday man, congrats on 40k too. well deserved

[Mino @compounded_mind]

@IterIntellectus a car nobody is forced to buy is a waste of money but destroying public property somehow isn’t

[Vlad Matsiiako @matsiiako]

If you're using OpenCode, here's what I'd do following today's breach:

1. Rotate potentially exposed API keys immediately. (@infisical can automate rotation for OpenAI, OpenRouter, etc.)
2. Stop handing raw API keys to AI agents. Use a transparent proxy like Infisical Agent Vault instead.
3. Use dynamic secrets so database credentials are short-lived rather than permanent.
4. Review audit logs for suspicious activity and unauthorized access.

[Matt Johansen @mattjay]

🚨 Confirmed. OpenCode hacked. Keys stolen.

Telling customers to revoke.

[Mino @compounded_mind]

https://t.co/tNhsHZRlLA

[Mino @compounded_mind]

people are already reporting hundreds of dollars in unauthorized API charges

did someone just use stolen OpenCode keys to distill frontier models at other people’s expense?

[OpenCode @opencode]

We had a security breach last night related to a large scale distillation attack that impacted 1.4% of OpenCode Console users

We have notified them all this morning via email. If you did not get an email you were not impacted, it was a subset of users who had external providers linked

The incident is ongoing so we cannot share more details at this point but we are working with Anthropic, OpenAI, and Google to remediate

[Mino @compounded_mind]

https://t.co/fjlQTuQwdT

[Mino @compounded_mind]

@opencode what exactly is the connection between the distillation attack and the API key exposure? were the stolen keys used to run the attack?

[Mino @compounded_mind]

@elonmusk insane

[Mino @compounded_mind]

@thsottiaux meanwhile on the other side of the card

[Mino @compounded_mind]

https://t.co/83Tbs9lSb2

[Mino @compounded_mind]

Anthropic just published a study on how AI is changing work inside the company

by August, its AI agents were running 6x as many hours as employees worked. the proportion of junior technical hires fell from 27% to 9%, while output per employee increased

one of the more interesting findings is that even nontechnical employees would need half their workweek just to read everything their agents produce

human attention is becoming the bottleneck

[Mino @compounded_mind]

https://t.co/R489ZaiiLU

[Mino @compounded_mind]

Claude was given 8 quantum circuit design problems and an automated evaluator to test its solutions

it ended up finding new designs that outperformed the best published ones across all 8 tasks, in some cases by more than 10x

one of the breakthroughs came from a 53-hour run with minimal human input

we’re getting to the point where AI can spend days independently working through difficult research problems and come back with something genuinely new

[Iam_ @SPAC89]

GPT 6.1 Sol Ultra vs DeepSeek V4.1 Flash 1017 Max, I still can't believe a model that's 90% cheaper can produce results this good, how is this even possible? Both models took roughly the same amount of time since most of it was spent rendering, but the quality DeepSeek managed to achieve is incredible, especially considering the API cost was just 0.44 cents for the entire test! I think having access to free web search is making a huge difference and giving DeepSeek a much bigger advantage than before, especially with its native multimodal capabilities. The more I test this model, the more I wonder how they're managing to offer so much at such a low price, I'm still running even bigger tests with different models to see how far it can go, more results coming soon!

[Mino @compounded_mind]

researchers found that neurons in the visual cortex rapidly alternate between representing what we’re consciously seeing and what we’re not, around 15 times per second, even when our perception stays stable.

really makes you wonder how much we’re missing in artificial neural networks by treating time as just another dimension of computation

what if the timing of neural activity is just as important as the information being processed?

[Doris Tsao @doristsao]

1/ We discovered something totally unexpected about the neural structure of consciousness that overturns decades of thinking on its biological basis and brings us significantly closer to the recipe for generating conscious awareness.
New preprint: https://t.co/K8BwiHXvXg (led by @HesseJanis, with @FrankLanfranchi, @_conor_mcgrory, @Yuelin_Shi, and @EngelTatiana)

[Mino @compounded_mind]

@deepseek_ai @DeepSeekHarness the whale delivered 🐋