Fábio Carvalho

🌀 Embark on a transformative journey with ‘Reflections of a Hypnotist’ by Fábio Carvalho. 🌟 Discover the power of hypnosis for positive change in this great Audible hit! #Hypnosis #MindPower #AudibleBooks 🎧✨ amzn.to/4aRX5Cb

AI's new phase isn't about chatbots. It's about agents inside the apps you already use.

@lorenzolfm Double spend se evita com ledger e transações dentro de lock. Reduza impacto de analytics com o @duckdb duckdb

Vc está construindo uma fintech, o banco é PostgreSQL. O que vc faz pra evitar double spend?

@mikefutia SORA

Sora 2 + n8n is absolutely insane 🤯 This n8n automation generates entire UGC campaigns with the same AI creator across unlimited videos. All from one Airtable form. Perfect for DTC brands & agencies who need brand consistency in their AI ads without hiring real creators. Why this matters: Every AI video tool gives you a random person each time. You can't build multi-video campaigns because your "creator" changes in every clip. Sora 2 consistent characters solves this: Same AI creator → Different scenes → Unlimited videos The n8n workflow: → Fill out Airtable form once (select your character, describe scenes, choose quantity) → Claude AI generates professional Sora 2 prompts automatically → Sora 2 renders videos with your consistent character → Videos auto-upload to ImageKit CDN → Everything tracked in Airtable with shareable URLs No manual prompting. No file management. No different people in every video. What you can create: → 3-part testimonial series with the same person → Before/during/after transformation campaigns → Product tutorial sequences that feel cohesive → Entire ad creative libraries with your "brand ambassador" Track everything in Airtable: → Video status (queued → generating → complete) → Shareable URLs for each clip → Scene descriptions and prompts → Production-ready in 5-10 minutes Built 100% in n8n + Airtable. Want the complete template? > Comment "SORA" > Like this post And I'll send it over (must be following so I can DM)

Habemus Papam 🙏

Obstacles are those frightful things you see when you take your eyes off your goal. — Henry Ford

R.I.P. Pope Francis

"Cada dia oferece seus próprios presentes." – Marco Aurélio #InspirarTransforma

@svpino @thedigitaldr How cool would it be if there was a GPT that would create the training dataset for fine tuning based our own custom data; as easy and simple as just uploading pdf and docs 🤓

@thedigitaldr It will choose from public HuggingFace datasets that match your problem. It's pretty cool. You can also provide your own dataset if you'd like. The system can help you with the format.

You can now fine-tune Llama 3 without writing a single line of code! We are moving at breakneck speed. I recorded a video to show you how to fine-tune any open-source model in a few minutes. I'm using a GPT capable of taking a problem and turning it into a fine-tuned model that will solve it. You don't have to write any code. You only need to explain to a GPT what problem you want to solve and tell it you want to use Llama 3. For example, "fine-tune Llama 3" or "deploy zephyr." It feels magic. The system will recommend a dataset and fine-tune the model for you. I'm using @monsterapis, a platform that specializes in making fine-tuning and deploying open-source models easy and fast. Their stack is well-optimized to maximize fine-tuning efficiency using techniques like Q-Lora and vLLM. They are behind the GPT. Here is what you need to do: 1. Create an account at monsterapi.ai 2. Load the GPT with the link below chat.openai.com/g/g-yWHAqw26c-… This is as simple as it gets. When you are done, you can click a button to deploy the model and start using it. I have 10,000 free credits for anyone using the code "SANTIAGO" in the monsterapi.ai/gpt dashboard. You can use these credits to access, fine-tune, and deploy these open-source models. You can also keep up with their latest updates, and get free credits and special offers on their Discord server: discord.com/invite/mVXfag4…

@Jilong123456 @Aurimas_Gr Critical steps on machine learning lifecycle

for continuous training, automation is key. here are some steps: 1. set clear triggers for retraining based on performance metrics. 2. ensure data pipelines are robust and can handle real-time data. 3. monitor model performance regularly to catch drift early. 4. implement version control for models to track changes. these steps help maintain model accuracy and relevance.

ML/LLMOps fundamentals - 𝗖𝗼𝗻𝘁𝗶𝗻𝘂𝗼𝘂𝘀 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 (𝗖𝗧) and what steps are needed to achieve it. CT is the process of automated ML Model retraining in Production Environments on a specific trigger. Let’s look into some prerequisites for this: 1) Automation of ML Pipelines. - Pipelines are orchestrated. - Each pipeline step is developed independently and is able to run on different technology stacks. - Pipelines are treated as a code artifact. ✅ You deploy Pipelines instead of Model Artifacts allowing Continuous Training In production. ✅ Reuse of components allows for rapid experimentation. 2) Introduction of strict Data and Model Validation steps in the ML Pipeline. - Data is validated before training the Model. If inconsistencies are found - Pipeline is aborted. - Model is validated after training. Only after it passes the validation is it handed over for deployment. ✅ Short circuits of the Pipeline allow for safe CT in production. 3) Introduction of ML Metadata Store. - Any Metadata related to ML artifact creation is tracked here. - We also track performance of the ML Model. ✅ Experiments become reproducible and comparable between each other. ✅ Model Registry acts as glue between training and deployment pipelines. 4) Different Pipeline triggers in production. - Ad-hoc. - Cron. - Reactive to Metrics produced in Model Monitoring System. - Arrival of New Data. ✅ This is where the Continuous Training is actually triggered. 5) Introduction of Feature Store (Optional). - Avoid work duplication when defining features. - Reduce risk of Training/Serving Skew. 𝗠𝘆 𝘁𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗼𝗻 𝗖𝗧: ➡️ Introduction of CT is not straightforward and you should approach it iteratively. The following could be good Quarterly Goals to set: - Experiment Tracking is extremely important at any level of ML Maturity and the least invasive in the process of ML Model training - I would start with ML Metadata Store introduction. - Orchestration of ML Pipelines is always a good idea, there are many tools supporting this (Airflow, Kubeflow, VertexAI etc.). If you are not doing it yet - grab this next, also make the validation steps part of this goal. - The need for a Feature Store will wary on the types of Models you are deploying. I would prioritise it if you have Models that perform Online predictions as it will help with avoiding Training/Serving Skew. - Don’t rush with Automated retraining. Ad-hoc and on-schedule will bring you a long way. Let me know your thoughts! 👇 #LLM #MachineLearning #AI

A verdadeira motivação não vem apenas da lógica do ‘porquê’, mas da emoção que nos move, e o primeiro passo para qualquer transformação é simplesmente decidir atravessar o limiar da ação.

@MarieMartens Problem-solving at the core 🎯 bootstrapping success

We did it 🥹 $2M ARR. Bootstrapped, with a full-time team of 5. It all started in 2020 when we asked ourselves: ❌ Why are forms so boring? ❌ Why are they so expensive? ❌ Why do they always look… bad? What if: ✅ Forms were actually fun to create? ✅ Forms had no volume-based pricing—unlimited submissions for free? ✅ We could build an independent company—no VC money, on our own terms? Fast forward to today, and I couldn’t be prouder to hit this milestone with @TallyForms. Our blog has almost become a personal diary, where we’re documenting every step of the way—and you can find the latest update here: blog.tally.so/we-crossed-2m-…

Point of view.

"The Singularity Is Nearer: When We Merge with AI" by Ray Kurzweil Kurzweil’s vision for a future where technology and humanity converge, offering insights into the possibilities and challenges that lie ahead. a.co/30bxxDK

Incredible to see AI advancing so quickly! Exciting times ahead!

Feeling nostalgic... AI helped me recreate a retro terminal experience! 😊 #AI #RetroTech

Fine-tuning isn’t just about tweaking numbers; it’s about understanding the model’s behavior under different configurations to extract maximum performance from pre-trained models.

Mastery over LLM model selection and optimization, such as fine-tuning and RAG, empowers you to align AI solutions with business goals and deliver measurable results.

DeepSeek R1, Janus… Kimi k1.5… MIT open source, reasoning, multi-modal LLM, scalable low cost… efficiency game race mode on #ai wow, what’s next?

@jazzapin @gregisenberg @crewAIInc @n8n_io @make_hq @FlowiseAI @langchain 🚀

@gregisenberg Where are current founders building these agents?

the next wave of startups won't launch with marketing teams they'll launch with AI agents running everything 24/7 - from content to acquisition to analytics. this isn't future talk. quietly there are founders who are printing with growth ai agents right now (but they won't tweet about it). imagine your acquisition agent runs 50 meme accounts simultaneously, testing hooks across different niches, generating 1000 posts daily until it finds what hits. your research agent analyzes 100k tweets per hour, finding unmet needs and feature requests that no one's building for. your content agent creates 200 unique hooks daily across X, linkedin, and tiktok, learning from each response, optimizing for what works. your community agent welcomes every new user personally, handles support tickets in seconds, and turns feedback into feature priorities. your SEO agent generates 500 pages of perfect content daily, while your ads agent tests 1000 creative variations across platforms, automatically killing what doesn't work. we already use this at @boringmarketer your email agent writes and tests 50 different sequences, personalizing every message based on user behavior. your analytics agent spots trends before humans could, adjusting strategy in real-time. the growth stack becomes fully automated, working 24/7. you get the idea. you pay per result. what previously required 20 people and $2M in salary now happens automatically with $2k in agent costs. one founder becomes as powerful as a funded startup. customer acquisition becomes predictable. growth becomes systematic. winners will be whoever has access to the best agent stack first, whoever has the taste to listen to the right signals and whoever can find the right niche at the right time. fun to think about because you can see it already start to happen. game on.

@MatInvest1 A vida acontece onde estiver nossa atenção.

Na década de 1980, meu primo imigrou para os EUA. Ele montou uma fábrica nos arredores de New Jersey. Se formos comparar, eu estou melhor que ele financeiramente. Porém, no final de semana, ele passeia no Central Park, vai no Metropolitan, leva os filhos no Museu de História Natural, caminhar em Long Island, nas férias leva as crianças pra acampar nos Apalaches. No sábado eu tomo uma cerveja no shopping, porque não há nada pra fazer nesse lugar 🤡 Quem se saiu melhor?