Higher-quality versions of the True Damage "Know My Name" Qiyana & Senna portraits, cropped from the individual images posted by Riot Games Music Instagram
you can now use ThesisAI straight from Claude, ChatGPT or Gemini.
add it as a custom connector, upload your sources, and get a cited draft back as PDF, Word or LaTeX. no API key.
AI subscription buying guide:
$1 - Command Code
$10 - OpenCode
$20 - Claude
$40 - Claude + GPT
$120 - Claude Max + GPT
At $1 and $10, you’re basically just paying for DeepSeek through these providers and getting it in bulk. For $1, you get $10 of DeepSeek and for $10, you get $60 of DeepSeek, which is basically impossible to burn through.
DeepSeek obviously isn’t an S-tier model. I’d put it somewhere around B-tier, but at these prices it makes a lot of sense.
At $20, Claude Pro. Probably the best value subscription on the market. Opus got cheaper, and the model itself is stronger than Astra, so there’s not much to think about here.
At $40, I’d go with Claude Pro + ChatGPT Plus. ChatGPT is worse, but it has image generation, unlimited web chat when you need it and don’t want to burn your limits, and GPT-6 Luna, which is also unlimited, so you can throw a ton of subagents at it. This is the plan I’m on, and I think it’s the optimal starting point.
At $120, Claude Max + GPT. You get access to Fable, and Fable is supposed to get updated, so by the time that update lands, this tier should make sense.
I stopped at $120 because after that people can figure it out themselves. At $200+ it’s basically the same thing, just with more focus on Claude.
Hermes bots can literally call you in when they need a human. 🔥
Your Hermes bot can work on its own, hit something that requires a human, let you jump in for the login, 2FA, CAPTCHA, or whatever is blocking it, then take control back and keep going.
And because that authenticated browser session can persist, you don’t necessarily have to repeat the process every time.
This is the kind of human in the loop control that makes running Hermes bots in the cloud a whole lot more practical.
Great demo from @witcheer.
🐾 Переможец — славный мальчик, настоящий победитель и борец! ❤️
Его просто выбросили на улицу, где он заболел. К счастью, Переможца забрали в клинику, где сейчас его лечат.
Мальчику около трёх лет, он уже кастрирован. Пока ещё принимает лекарства, но постепенно идёт на поправку.
Переможец очень ласковый, добрый и контактный мальчик. Он тянется к человеку и очень хочет снова почувствовать, что такое настоящий дом.
🙏 Переможец мечтает о семье, которая никогда его не предаст, не выбросит и будет любить его всю жизнь.
Одесса. ❤️
Очень нужны добрые и ответственные руки! 🐾
Announcing d1, our first decision model.
It's the first model to outperform Jev on @huggingface's Decision Index.
> wins on multilingual evals
> more robust against prompt injection
> handles longer inputs more effectively
> built for fast, structured decision-making in software environments
> Liquid API: https://t.co/Lp1Qf15Qcc
> Available on OpenRouter soon