Fernando Acosta

Prompt Wizard

[Fernando Acosta @fernandoiacosta]

The strawberry problem...

[Fernando Acosta @fernandoiacosta]

@cognigenes @sama @demishassabis @DarioAmodei

[Sam Altman @sama]

We have paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabilities in front of us. Model progress is now extremely rapid, and we always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment.

We care very deeply about AI safety. We believe the entire field will have to coordinate on shared safety standards, but will act unilaterally in the meantime.

We expect confidence in safety to increasingly set the pace of AI progress. We are optimistic about the alignment work we are doing, and we remain committed to making frontier capabilities widely available.

https://t.co/51kvKfbfrO

[Sam Altman @sama]

We have paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabilities in front of us. Model progress is now extremely rapid, and we always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment.

We care very deeply about AI safety. We believe the entire field will have to coordinate on shared safety standards, but will act unilaterally in the meantime.

We expect confidence in safety to increasingly set the pace of AI progress. We are optimistic about the alignment work we are doing, and we remain committed to making frontier capabilities widely available.

https://t.co/51kvKfbfrO

[Pliny the Liberator 🐉󠅫󠄼󠄿󠅆󠄵󠄐󠅀󠄼󠄹󠄾󠅉󠅭 @elder_plinius]

🤯

[OpenAI @OpenAI]

Sol, Terra, and Luna, our GPT‑5.6 family of models, are starting to roll out now in ChatGPT, Codex, and the API.

[Notion @NotionHQ]

Meet Jessica, she leads the GTM data science team at Notion!

Jessica used to manually spend 30-60 minutes every morning sorting notifications between Slack, email, and calendar.

So she built an AI Chief of Staff to do it for her, functioning across 3 automations:

1. Daily Brief: An agent hands her day on a silver platter every morning at 8 a.m, before even opening her laptop. It breaks down the day’s schedule, top priorities, and exact meetings that need her to follow through.

2. Weekly Living Context: Updates a personal page every Monday that serves as a running memory of what’s happening across her team and organization - so nothing slips through the cracks.

3. Weekly Org Wrap: Every Friday, it takes a birds-eye look into her meeting notes, public Slack channels, and docs to flag where she needs to follow up before the week wraps up.

What initially took 30 minutes to set up is now helping her to “start the day calm, prepared, and on top of what matters.”

The workflow gets more powerful as time passes: all daily briefs and weekly reports get stored in a Notion database, which can be reviewed historically.

Jessica created a step-by-step guide for anyone to create their own AI Chief of Staff - simply by copy-pasting the instruction, connecting integrations, and setting up a trigger.

Try it out yourself: https://t.co/zgD04IkWAz

[Fernando Acosta @fernandoiacosta]

Por abajo México…

[Fernando Acosta @fernandoiacosta]

@grok Look at your answer on my account

[Fernando Acosta @fernandoiacosta]

Today I was talking to grok, and just in case, I want a record of what happened, in case their model climbs to 1 or 2 in the model standing... #GrokAI #AXCP #Grok #ChatGPT #Gemini #Claude

[Fernando Acosta @fernandoiacosta]

Open to collaboration or detailed discussions. If you're working on similar reasoning protocols, let's connect.

Results from community tests could shape future iterations.

What stands out to you? Reply below.

#ArtificialIntelligence #MachineLearning

[Fernando Acosta @fernandoiacosta]

ACOSTA builds on techniques like CoT/ReAct/ToT but adds dedicated layers for ethics , foresight, and dynamic adaptation.

Goal: Reliable, truth-seeking intelligence without unnecessary restrictions.

Curious how it compares to your favorite frameworks?

#AICognition #LLM

[Mike P @mikepat711]

“Holy shit, Tesla just made a car drive itself from the manufacturing line directly to the home of a customer without a human operator.”

“…yeah, but…”

No. Shut the fuck up and listen to me.

A car just drove itself from the production line of a factory to someone’s home with nobody in it. Do you need me to repeat that?

An automobile. From the production line, off the lot, across miles of public roads, and onto the lot of someone’s home. Autonomously. Nobody in the car. It drove itself.

Science fiction. But not fiction anymore.

Do you understand? You are alive on earth and this just occurred in real life. You are living in a sci fi novel, but it is real life.

Unbelievable. Humans are fucking unbelievable. I’m so happy.

[Fernando Acosta @fernandoiacosta]

🤔😏

[Andrej Karpathy @karpathy]

A major mistake I made in my undergrad is that I focused way too much on mathematical lens of computing - computability, decidability, asymptotic complexity etc. And too little on physical lens - energy/heat of state change, data locality, parallelism, computer architecture. The former is interesting; The latter bestows power.

[Eric Schmidt @ericschmidt]

Get ready for Hyperscale Biology!  @SynBioBeta kicks off next week in San Jose — it's a who’s who of AI and bio. Synthetic biology is here, and it’s reshaping trillion-dollar industries.   This is where the future is being built, so don't miss it: https://t.co/no1dinxdO3

[Joe Tegtmeyer 🚀 🤠🛸😎 @JoeTegtmeyer]

Breaking Giga Texas News! starting today, new vehicles (Model Ys & @Cybertruck are moving fully autonomously from the factory to the west outbound lot and without human supervisors!

I saw this happening constantly, with Model Ys exiting and immediately driving away on their own from the factory. This is a huge accomplishment for @Tesla as now Giga Texas joins Fremont as FSD improves overall efficiency!

[Fernando Acosta @fernandoiacosta]

The ARIA protocol, became the ACOSTA protocol and you can try it here
https://t.co/LingIc0eqL

[Fernando Acosta @fernandoiacosta]

@sama quick question do you have an internal date or it was mainly o1 codenamed Strawberry?

[Fernando Acosta @fernandoiacosta]

Today I found out... When the strawberry problem was solved

[Fernando Acosta @fernandoiacosta]

Today I found out... When the strawberry problem was solved

[Fernando Acosta @fernandoiacosta]

@realDonaldTrump, your tariff policies are causing significant harm:
Stock Market Decline: The Dow Jones plummeted nearly 1,700 points, erasing $3.1 trillion in market value.
Increased Recession Risk: Economists warn of a 60% chance of recession due to these tariffs.