Edward Alexis

Founder of

[Edward Alexis @edalvb]

@Da7_Tech Yo sentí eso al segundo día del lanzamiento y además que consumía más tokens.

[camenduru @camenduru]

👾🎱 Turn videos into game-ready sprite sheets (8-direction). 100% Local. No external APIs.

https://t.co/fcNwcSzsMh
https://t.co/HisshkovYl

Thanks to ❤

https://t.co/404K8pL368
https://t.co/lRSQJgDdgh
https://t.co/1Pm8lsDsdO
https://t.co/nP78fuXRDY
https://t.co/qTt3W2ZELW
https://t.co/yIYAwNh6xU
https://t.co/6TAQRTej1v
https://t.co/kyZuj6JrQg

[Da7em @Da7_Tech]

I told Fable how much I love its style, how it works with me, and how well it gets me, and that I needed that in Hermes Agent.

Now Fable's vibe and my preferences are baked into Hermes Agent.
Whatever other model I run, I get way better output because Fable's mindset drives it.

[Nils @broodsugar]

We can measure what part of the shelf gets customer attention without any additional hardware. Just your normal CCTV cameras and a digital twin made with the phone in your pocket.

@CactusXR will absolutely revolutionize how well stores can do their merchandising.

[Fazt @FaztTech]

Lo más difícil de desarrollar software con IA para negocios reales, no es pedirle que cree una API, diseñe o haga frontend + backend + base de datos. Eso la IA ya lo puede generar.

Lo difícil es entender al cliente: no siempre sabe lo que quiere exactamente, su problema suele ser más complejo de lo que cree y, si por él fuera, te pediría una interfaz con 50 botones y una tabla enorme al estilo Excel, sin mencionar que no tiene ninguna consideracion de como escalar un sistema real.

Ahí está el trabajo real del desarrollo actual: ser el intermediario entre la IA (la que genera el código) y el cliente (el que explica el problema).

[Edward Alexis @edalvb]

@Da7_Tech En la 7 la cagué hace unos meses, estuve trabajando con Claude, me pasé a Codex para "probar" me jodió todo el proyecto, ahora ya resolví los problemas y sigo con Claude. 💡

[Da7em @Da7_Tech]

I ran a detailed comparison between Fable 5.1, Astra, and SWE-2 in Devin to map out their strengths and weaknesses:

SWE-2

– Strengths: It explores territory other models never touch, looking at problems with a much wider lens. It picks up on subtle cues with very little context. You can trust it with minimal prompting, like telling it to delete something without worrying it might wipe out unrelated code. It has great discipline. When evaluating a problem, it targets the main pain point right away instead of wasting time on trivia. Its "what not to do" advice is genuinely useful and keeps you from spinning your wheels in solved areas. Unlike Astra, it avoids arbitrary, nonsensical refusals. It feels much more honest about what it can and cannot do.

– Weaknesses: It lags far behind its base model, Kimi K3, in UI sense, aesthetics, and design taste. It also tends to stumble on a great idea and then hesitate to actually ship it.

Astra

– Strengths: Pure tenacity once it identifies a bug. It simply doesn't stop until it solves the problem. In one test, Opus 5 gave up and told me to file a support ticket, while Astra kept digging until it fixed the issue completely. It shows real engineering creativity, picking tight research paths with concise, structurally sound solutions and solid recommendations. It also has a great sense of success probability and offers precise statistical qualification for tasks.

– Weaknesses: It tends to plan around small sample sizes and thin evidence. It gets bogged down in minor details instead of looking at the big picture and overall impact. It also struggles to synthesize ideas across different domains, overspecializing until the scope becomes too narrow.

Fable 5.1

– Strengths: The highest evidence accuracy by far. It reliably delivers on what it promises, unlike models that promise everything and underdeliver. It has strong design taste, clear recommendations, and a holistic grasp of the end goal. It weighs priorities accurately. It can also pivot your entire strategy convincingly without drifting off into irrelevant territory.

– Weaknesses: Its answers are dense and exhausting to read. Once it settles on an idea, it's hard to steer it away. Its research skims the surface rather than digging up rare findings. It can suffer from tunnel vision, sticking strictly to technical aspects while ignoring costs or viability metrics unless prompted.
Patterns and Behaviors

– Negation: Fable and SWE-2 set broad boundaries. Astra uses rigid negative rules that abruptly flip decisions when triggered.

– Stability: Fable and Astra are steady. SWE-2 can be inconsistent, sometimes generating an idea only to backtrack on it.

– Length: Astra is the most concise. Fable is the longest. SWE-2 sits in the middle, though it has repetitive filler that needs trimming.

– Main Risks: SWE-2 over-explores possibilities at the expense of verifying existing code. Astra loses the forest for the trees. Fable turns practical technical decisions into academic literature reviews.

– Standout Traits: SWE-2 wins on thoroughness, Astra wins on grit, and Fable wins on raw reasoning and persuasion.
Scores and Recommendations

– Scores:

Fable 5.1: 8.5/10.
Astra: 7.5/10.
SWE-2: 7/10.

– Pick Fable or SWE-2 if you want strict prompt adherence, clean execution, or deep mechanical logic.

– Pick Astra if you need precise verification evidence and relentless debugging.

– Scope and context: SWE-2 and Fable are tied, while Astra falls slightly behind.

This is an initial assessment based on tests in Devin, and I'll need more runs to map out the nuances completely.

[Miguel Ángel Durán @midudev]

¡Cloudflare convierte tu agente de IA en un auditor de seguridad! Esta skill es la que usan ellos internamente.

Lo hace en 6 fases: reconocimiento, caza, validación, verificación, salida estructurada y reporte.

Código abierto:
→ https://t.co/wv4OWasFDv

[Kamran Ahmed @nilbuild]

Made a skill for interactive mascots

/𝚙𝚊𝚐𝚎-𝚖𝚊𝚜𝚌𝚘𝚝 [𝚢𝚘𝚞𝚛 𝚌𝚑𝚊𝚛𝚊𝚌𝚝𝚎𝚛]

It generates a mascot that follows your cursor and reacts when you poke it

https://t.co/exU2rTVh0M

[Fazt @FaztTech]

Si necesitas un diagrama de arquitectura y no te gusta lo que genera Mermaid o una imagen con IA, conoce Archify.

Es un skill para Agentes como Cursor, Claude o Codex: te genera un HTML interactivo con el tipo de diagrama que pidas.

Sirve para explicar un sistema, dejarlo en la documentación o exportarlo a PNG y otros formatos.

https://t.co/CRPZ5xKlqP

[ashe @ashebytes]

since you guys loved the exploding tesla..

I used GPT-6 Astra to create a 3D website that pulls apart the male anatomy into 2,234 modeled pieces!

we are in a renaissance of learning

[Da7em @Da7_Tech]

I get why regular people try out models and call them AGI. But when OpenAI's founder publicly claims Astra reached AGI, that's a disaster. It shows he doesn't even get what AGI means.

AGI is simple. It's a model that can solve problems it was never trained on, matching frontier human cognition by coming up with brand-new solutions.

Right now, everyone hyping Astra as AGI keeps pointing to computer use. Didn't we already have that months ago with GPT 5.5, qua driver, and other tools? Sure, it's faster. Does that make it AGI?

People also rave about its 3D output. Yeah, it's impressive, I'll give it that. But it's still just training on known samples, workflows, and data where it happened to beat the competition. That's still not AGI.

Fable is still way better when it comes to taste, general understanding, awareness, and holding context.

Fable actually challenges your imagination and pushes you to places you hadn't thought of. Astra hasn't given me that feeling at all.

Let's be real. It's much better than Sol, and it's a huge leap for OpenAI, but marketing it as AGI insults people's intelligence.

Current models are great at inferring from what already exists. Creating something genuinely new from scratch? Only one lab has managed that internally, and we won't see that model for at least four or five years because it's far from ready.

Get one thing straight: true AGI or ASI will never be handed to the public through a simple subscription. There will be strict regulations, safety limits, and heavy nerfing. Don't buy into the delusion that you'll run AGI or ASI at home, locally, or through a $200 subscription.

Anyone telling you that is either clueless or lying straight to your face.

The reality is, once a model hits the ASI stage, no lab will be able to contain it. It'll be a massive danger to humanity, and what scares me most is that we're heading right toward it.

[Edward Alexis @edalvb]

@Da7_Tech Totalmente de acuerdo.

[Da7em @Da7_Tech]

Want to see the real difference between Fable and Astra?

Give both a massive task with almost zero guidance, then check their work.

Astra will pull moves that drive you crazy and make you regret not micromanaging it.

Fable, on the other hand, nails the judgment calls. It feels less like an AI and more like an absolute top-tier human making decisions you'll actually be thrilled with.

[Claude @claudeai]

We’re introducing Claude Fable 5.1 and Claude Mythos 5.1.

They're the world’s most advanced models for coding and knowledge work.

[Da7em @Da7_Tech]

Devin vs Cursor on the $200 plan:

Want simplicity? Cursor.

Want control and customization? Devin.

If you mostly stick to first-party models, Cursor is the better deal.
Grok 4.6 runs circles around SWE-1.7, and the monthly limit is generous enough that you won’t think about it.

If you lean on third-party models like Fable, Sol, and Kimi, go with Devin.

The weekly limits there are far more forgiving, and a few of those models cost you nothing.

You can run Fable and Sol all week on Devin. On Cursor that’s simply not happening.

Me personally? I’m going with Devin.

[Da7em @Da7_Tech]

The two skills that matter most right now, especially with AI and how competitive everything got: reading and writing.

Look around. Almost nobody reads. Put a long piece in front of them with ideas, a narrative, some soul, a lived experience, and they immediately want the summary. They ask Grok to crush it. They don’t want the human experience. They want the lifeless extract that leaves no mark.

I looked into this myself. It’s a marker of mental decline. Attention is already wrecked by short clips, addictive apps, doomscrolling.

If you can’t absorb a human experience, you can’t generate one.

You can’t give what you don’t have. You won’t speak well. You won’t write well. You won’t direct agents well, deal with people well, say what you actually think, or make anything that stands out.

Copy a method for using agents and you copied a method. Understand something from an actual life and you can invent. That’s the difference.

Read whatever you want. Literature, history, philosophy, anything. The more range your mind has, the more imagination you have. The more imagination, the more you can actually make something great with agents instead of running someone else’s workflow.

Read. Learn. Talk. Write. Share.

That’s the sweetness of the human experience. No AI replaces that.

[Beatriz Aparicio @VeritatisPeru26]

Que vuelva esta ley 895 derogada por Toledo
@KeikoFujimori

[Google Research @GoogleResearch]

Introducing GlucoFM, a lightweight, self-supervised continuous glucose monitoring foundation model that separates metabolic baselines from transient spikes, producing transferable representations and setting new performance standards across diverse metabolic prediction tasks, such as diabetes risk assessment, insulin resistance, and post-prandial glycemic response.

Learn more →https://t.co/PA7wpFz2xo

[Peter H. Diamandis, MD @PeterDiamandis]

A robot outran Usain Bolt. A personalized cancer vaccine passed Phase 3. A virtual human cell went live. A million daily drone deliveries got announced. That was ONE week. What's the most underrated story I missed?