Ejercicio del Episodio 01.
Spotify bit.ly/4zgIJGF
YouTube youtu.be/2aIf9LPvwXw
#droso101 #podcast #economiadelcomportamiento #IA #productividad
🤖 Made with AI
droso101 retweeted
i reverse engineered wispr flow, then used deepseek to fine-tune LFM3-1.7B. the result is on par with the wispr flow models you pay for, and it's free, open source and runs entirely on your mac.
key-up latency is 65 ms on my m4 max vs ~520 ms for wispr flow on the same clips. a 6 minute yap session pacing around my office still lands in about a quarter second.
> rust backend with a native metal engine: each decode step runs as one fused megakernel
> a dflash2 speculative drafter is built into the engine, trained on my local 4 x rtx pro 6000 server
> parakeet transcribes, LFM3-1.7B refines. both stream while you talk, so key-up only waits on the last chunk
> never leaves your macbook, so slow internet or bad ping is no longer an issue
> import your own dictionary of terms for your workflow
> i picked every default myself. it feels amazing out of the box
> adapts to the app you're in: discord is lowercase casual, slack and email are formal, claude code gets clear prompts
> training data: 77k transcripts of public audio and synthetic dictations, each rewritten by deepseek v4.1 flash 1017 following a written spec of what "what you meant" means, in 4 writing modes. none of my own dictations went into it
> blind judge (claude opus) on an hour of lectures: phonon's text preferred 90 times, wispr's 23
> built with claude opus 5.5 [xhigh]
brew install --cask infatoshi/tap/phonon
Fluidez no es verdad.
Spotify bit.ly/4zgIJGF
YouTube youtu.be/2aIf9LPvwXw
#droso101 #podcast #economiadelcomportamiento #IA #productividad
🤖 Made with AI
droso101 retweeted
We want Antigravity to be working for you all the time. We’ve been rolling out remote agents this week. Your tasks run in independent cloud virtual machines. Fire off a task, step away for coffee, unblock permissions from your phone, and come back to passed tests and a clean pull request. Available on all paid consumer plans.
droso101 retweeted
La comunidad de IA educativa ya está en Instagram: ChatGPT-IA-edu, Vibe Coding Educativo y EduMedia-IAG, sus boletines y el Diario de IA Educativa. Cada mañana, el número del día; cada domingo, la semana en vídeo.
👉 Instagram: instagram.com/comunidad.iaed…
#IAEducativa #EdTech
droso101 retweeted
Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment.
The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work.
GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time.
openai.com/index/introducing…
droso101 retweeted
Hermes bots have coworkers now. 🤖🔥
HERMES BOT GROUP CHATS NOW RUN THEMSELVES.
Drop a message that names nobody and the room picks who takes it. "Assigned to Builder" shows up before anyone replies. Out now in Hermes v0.23.0.
What it does:
▸ The room picks who answers, or you give it a lead bot that hands out the work
▸ Bots pull each other in, and you see every handoff
▸ Each bot shows what it's doing live, like "scout · terminal · 7s"
▸ Scheduled jobs post into the room and wake only the bots they @mention
▸ Every room gets a purpose and shared notes each bot reads on its turn
▸ Bots from different machines can share one room
▸ Long discussions pause at the limit. Hit Continue to keep going
What you can do with it:
▸ Post a bug and let the room hand it to your builder bot
▸ Have a morning job drop the news in the room and wake only your writer bot
▸ Put your project brief in the shared notes so every bot works from the same plan
▸ Let a scout, a builder and a critic talk a plan through, then hit Continue if they need more rounds
Your bots stop being a row of separate chats. They work like a team now.
ALT Hermes v0.23.0 infographic: Your bots work like a team. Before, a row of separate bot chats (Builder, Scout, Writer). Now, one room with Builder, Scout, Writer and Critic, tagged 'Assigned to Builder'. Below, four use cases for Hermes bot group chats.
droso101 retweeted
🤯THESE 9 FREE ALTERNATIVES WILL SURPRISE YOU!
1. perplexity.ai — Search the web and get AI-powered answers.
2. bing.com — Find answers with AI-powered search.
3. gemini.google.com — Write, research, and brainstorm with AI.
Save this post
Quién eres, para quién es, qué quieres.
Spotify bit.ly/4zgIJGF
YouTube youtu.be/2aIf9LPvwXw
#droso101 #podcast #economiadelcomportamiento #IA #productividad
🤖 Made with AI
droso101 retweeted
Dejo esto por aquí y me voy a dormir que ha sido un día intenso de trabajo.
markitown.app/es-es/
droso101 retweeted
📌 RULES FOR CHARISMA | BECOMING MAGNETIC IN ANY ROOM
🔸️ Be Fully Present → Give people your undivided attention.
🔸️ Make Eye Contact → Show confidence and genuine interest.
🔸️ Listen More Than You Speak → Make others feel heard and valued.
🔸️ Use Their Names → Create a more personal connection.
🔸️ Speak with Purpose → Be clear, calm, and intentional.
🔸️ Bring Positive Energy → Make people feel comfortable around you.
🔸️ Show Genuine Interest → Ask thoughtful questions instead of trying to impress.
🔸️ Be Comfortable with Yourself → Confidence is magnetic when it doesn't need approval.
Charisma isn't about being the loudest person in the room. It's about making others feel good in your presence. 💪
📩 Free Guide: Master the 12 soft skills that drive success—and start applying them today. 🚀
👉 Get your free copy:
betteryoubrief.substack.com/…
#Charisma #SocialSkills #PersonalGrowth
droso101 retweeted
Most people learn Claude in random order, then hit a ceiling because they skipped a layer. This map lays out seven stages from basic prompts to AI employees, and the order matters more than the count.
Foundation comes first. Learn what Claude can do and where it fails, then structured prompting with role, context, goal, constraints, and output. Add context engineering, which is giving Claude the right information before you ask, plus few-shot examples and styles that match your writing and brand.
Workflow is where you stop starting from zero. Projects save files, instructions, and knowledge together. Prompt chaining breaks complex work into search, analyze, create, review, and refine. Deep research, data analysis, and artifacts turn Claude into a research partner that reads your files and builds editable docs and dashboards.
Power features connect it to your world. Skills are reusable workflows Claude loads on its own, connectors and MCP link external tools and data, Claude Code works with your real codebase, and CLAUDE.md holds project-level instructions that persist.
Agents are where Claude stops waiting for you. It decides when to call tools, runs goal, plan, act, review loops, delegates to subagents, and keeps a human in the loop for approval. Claude Code mastery goes deeper into hooks, plugins, Git and GitHub workflows, testing and QA agents, and multi-agent development.
Then you build AI employees. Each one has a role, like research, content, sales, customer support, data, finance, or developer, and a defined job. The last stage turns those into systems, products, and business workflows.
The path runs prompt, project, artifact, skills, connectors, agents, AI employees. Same tools, bigger possibilities, and you only move up by finishing the step below.
Bookmark this before you try to build an agent while you are still on beginner prompts.
droso101 retweeted
Google isn’t trying to win the AI race.
They’re trying to own the entire AI Agent ecosystem.
While everyone argues ChatGPT vs Claude, Google quietly built:
Models → Gemini Pro, Flash, Deep Think, Gemma
Design → Stitch, Whisk, Imagen
Research → NotebookLM, AI Mode
Video → Veo, Flow, Google Vids
Coding → Antigravity IDE, Gemini CLI, Jules
Agents → A2A, ADK, FileSearch API
The scary part?
All of these tools talk to each other.
That means:
10x faster prototypes
End-to-end AI workflows
Production-ready agents on GCP
The next AI war won’t be model vs model.
It’ll be ecosystem vs ecosystem.
Save. Share. Build.
droso101 retweeted
📰 «En la Politécnica de Valencia van a poner arcos a la entrada de las aulas de examen. Como en los aeropuertos». 👉 profesorproductivo.com/blog/… @iayeducacion
La confusión de categoría explica la mayoría de las peleas con la IA.
Spotify bit.ly/4zgIJGF
YouTube youtu.be/2aIf9LPvwXw
#droso101 #podcast #economiadelcomportamiento #IA #productividad
🤖 Made with AI
droso101 retweeted
If your AI characters sound fake, the fix is almost always in how you write and DIRECT the dialogue in your prompt.
I recently started on the dialogue-heavy scenes for my AI feature film about Beethoven, and too many of the lines kept coming out wrong.
They’d sound too theatrical, too monotone, or too much like a practice read-through between actors.
(all this despite the fact that Seedance 2.5 is fantastic when it comes to realism.)
So through a whole lot of trial and error, I've found that these 5 things made the biggest impact on the quality of the dialogue delivery in my generations:
1. Paying close attention to the written punctuation used.
2. Breaking long lines into shorter ones, with intentional pauses.
3. Using bodily actions as punctuation (this one gave the biggest improvement BY FAR! In the attached clip, note the sniffles in between Beethoven's lines. That alone made the delivery feel much more natural.)
4. Using "anchor gestures" (he lifts his finger as if to say "one sec")
5. Just split your generations in two (stupid simple, but it really makes a difference.)
So if your characters keep sounding like they're reading from a teleprompter, give this a try!