Apoca

n8n + LLM workflows that survive production. Co-founder & AI dev at @PromptDimension. Real builds, real bugs, AI news that matters to builders.

[Apoca @Ap0cah0lics]

Ahahahha, no way, you see it???

Mates, fellas and all remarkable dudes, take this prompt and do something amazing!

[Prompt Dimension @PromptDimension]

Yes, now you can have your own dot via ChatGPT!

Greetings, trash pandas!🦝

We’ve created a prompt to help you generate your own "dot" based on your logo. Admit it—you all loved it.

A quick note on the prompt!
We’ve included a {{MEME}} placeholder where you can insert any meme face you’d like to feature on your dot.

P.S. We went with the Lenny Face.

[Apoca @Ap0cah0lics]

@icanvardar Prove it!

[Apoca @Ap0cah0lics]

@Tech_girl If you are talking about local hosted, then you are hosting an already trained model that has already "established patterns" and a certain basic understanding. Especially strong models therefore need a powerful system, because they require high computing power.

[Apoca @Ap0cah0lics]

@ai_for_success So far, there is nothing special, because only 1 dot agent is available. In general, it is no different from the usual Hermes or custom Claude. We need to wait until they give the opportunity to work with several dots at the same time.

[Apoca @Ap0cah0lics]

@solopribuilds 100%

[Apoca @Ap0cah0lics]

@theslowtell So far, I analyze manually and decide what to do next. Having collected the base a little, I plan to involve Claude in this business.

[Apoca @Ap0cah0lics]

@ai_for_success The latency improvement is cool, but a 0.2% WER difference is a tie. In the real world with noise and accents, these lab tests don't mean much.

[Apoca @Ap0cah0lics]

@J_srv001 Managing timelines declaratively with React is just so much easier than drawing frames manually in Python.

[Apoca @Ap0cah0lics]

@theslowtell Preparation of posts, tracking of replays, planning of posts and their automatic publication, calculation of analytics.

Plans to integrate the following TikTok and video generation using appropriate LLM models.

[Nous Research @NousResearch]

Hermes is for the wizards

[Apoca @Ap0cah0lics]

@NousResearch Ahahaha, this is very hilarious 🤣

[Apoca @Ap0cah0lics]

@pcshipp Share something interesting what did you do with opus 5.5?

[Apoca @Ap0cah0lics]

@adxtyahq Yes, indeed, in some cases this will speed up the process, but unfortunately this is not a panacea for the legal case in general.

[Apoca @Ap0cah0lics]

@danellisona The five crypto bots that replied to this in under a second definitely agree.

Ops, I think I'm talking about myself...

[Apoca @Ap0cah0lics]

@kimmonismus It's wild that a mid-tier model caught up to the flagship in just a few months.

[Apoca @Ap0cah0lics]

@CompleteSkeptic Do you already have any ideas where jev can be used in automation?

[Apoca @Ap0cah0lics]

@TeddyinMedia E2B is such a lifesaver for sandboxing. Setting that up from scratch is a massive pain.

[Apoca @Ap0cah0lics]

@rezoundous Yeah, coding with Claude lately has felt way ahead of everything else.