Ronald Ashri

Understanding AI Agents through the Generality - Accuracy - Simplicity (GAS) Framework Understanding and clearly articulating challenges and trade-offs is key to successful AI Agent deployments. A recent paper provides a great framework through which to express them.

understanding and clearly articulating challenges in AI agent deployment is key - a recent paper provides a useful framework through which to express them

open.substack.com/pub/agentsde...

great summary of what LLMs are capable of 😂

How a Simple Question Exposes Claude 4’s Shortcomings: "Claude, can you write a game?" The answer, my friends, is "no" - but Claude 4 doesn't know it and Anthropic chooses to leave things this way.

open.substack.com/pub/agentsde...

Super-Assistants Vs The World: Why LLM Wrappers, traditional SaaS tools and Specialist AI Agents all need to worry. The big LLM companies are launching do-everything super assistants. Where does that leave LLM Wrappers, specialist AI Agents and traditional SaaS tools?

Super-assistant like Claude and ChatGPT are coming after everyone

open.substack.com/pub/agentsde...

Very disappointing.

CIAO Design Principle #2: Explicit Rules in Code Beat Implicit Rules in Prompts CIAO (Conversations-Interfaces-Agents-Orchestration) is an AI application framework that is not afraid to clearly say that explicit code is better than prompts.

CIAO Design Principle #2: Explicit rules defined in code are better than implicit rules defined in prompts.

open.substack.com/pub/agentsde...

Il discorso di papa Leone XIV ai cardinali: “Il nome scelto per rispondere alle sfide dell’IA” Prevost ha incontrato i porporati per un saluto finale e per condividere alcune riflessioni: “Mi aiutate in giogo superiore alle mie forze”

Pope Leo the 13th advocated for workers rights and fair wages during the first industrial revolution.

The new Pope - Leo the 14th - apparently picked the name to “respond to another industrial revolution and to the developments in artificial intelligence”.

www.repubblica.it/cronaca/doss...

CIAO 👋 - A Unifying Architectural Framework for AI-powered applications We are creating a new type of software application.

CIAO stands for Conversations - Interfaces - Agents - Orchestration and it is an architectural framework for describing AI-powered applications to help teams better reason about what they actually need to design and build

open.substack.com/pub/agentsde...

Carl Sagan on how humanity can create its purpose and make its existence meaningful.

YouTube video by Pulp Pulp - Spike Island (Official Video)

“Maybe we need to work on other ways of coming alive”. youtu.be/-27a1ugJX8U?...

1. LLM-generated code tries to run code from online software packages. Which is normal but
2. The packages don’t exist. Which would normally cause an error but
3. Nefarious people have made malware under the package names that LLMs make up most often. So
4. Now the LLM code points to malware.

scooby doo meme revealing APIs hidden under the agent 2 agent protocol

[Emerging Standards] A2A Protocol: New Era of Agent Collaboration or Just APIs in a Trench Coat?

open.substack.com/pub/agentsde...

In comparison, what Argentina’s Milei did with crypto was amateur time. What are mere 10s of millions of a crypto scam compared to the 100s of millions of pump and dump currently going on in the USA.

The full report by the Tony Blair Institute is out, in which they detail how copyright should be "rebooted" in the AI age. It reads like a big tech lobbying document. I hope they will be asked serious questions at its launch in London today.

Some very brief highlights:

🧵 1/8

Disaster 🙂 It is a great illustration of how they are just “playing the odds” with no understanding of anything they are saying. Doing it with perfect confidence as well!

Scrolling through #TeslaTakeDown demonstration postings across the world is a great Sunday morning pick-me-up. Highly recommended.

"When this government says disabled people ought to work, we ask:

Where are the jobs?

The level-access commuter trains?

The accessible homes?

The carers to dress me at 6am? (Right, Brexit made them leave.)

The reduction in wait times for Access to Work, Wheelchair Services?"- Anna Landre

🤦🏽‍♂️ know what’s a really great way to learn how to think, break down problems and communicate clearly? Learn how to CODE!

come on who hasn’t accidentally cc’ed someone at work and then went on to share highly classified information regarding an imminent military strike while also badmouthing their allies and attempting to set a price for their bombing service while also gloating about how well they killed people

AI Is Not Your Friend Your computer does not care about you.

open.substack.com/pub/jurgengr...

The keyset is an input device with five long, flat keys, similar to piano keys but thinner. The device and the keys are beige. A braided wire cord exits the device at the top.

An image from Engelbart's demo. At the top, the screen shows lines of text. The bottom of the image shows Engelbart at a special desk. The desk has a keyboard in the center with a keyset on the left and a mouse on the right.

In 1968, Douglas Engelbart revolutionized computing by showing the mouse, hypertext, and collaborative computing in "The Mother of All Demos". Most of his ideas are now part of everyday computing, but one device—the keyset—failed to catch on. Let's look at Engelbart's demo and the keyset. 1/N

Given the amount of hype and the amount of actual useful projects in the real world I’d say that is probably about right. That 10% is still really interesting but the 90% is deafening.

I'm sorry but Jasmine Crockett saying, F"ind a little trans child that is ruining your life!" is well giving me life. She is both very right and very funny.

I helped build a government AI system. DOGE fired me, rolled the AI out to the whole agency, and implied the AI can do my job and the jobs of the others they've fired.

It can't. But, what DOGE accidentally revealed about themselves in the process is fascinating. 🧵

One mother’s win over Meta will change social media for everyone Britons will be able to opt out of targeted advertising after Tanya O’Carroll’s David-and-Goliath battle with Meta

Victory for @tanyaocarroll.bsky.social in case against #Meta!

Meta affirms right to OPT OUT of surveillance advertising

Another blow to the unethical and unfair business models of Meta, Google and X: you do not have to put up with targeted ads and profiling

www.thetimes.com/uk/technolog...

[AI Agent Diaries] The LLM industry's dirty secret and what to do about it. The LLM industry is experiencing a mini ChatGPT-in-reverse moment this week with The Atlantic's release of the LibGen search tool, showing how much pirated material is used for LLM training.

What the LLM industry is doing is not dissimilar to what robber barons across the ages have done. Sam Altman, Mark Zuckerberg and others have identified a natural resource, content produced over centuries of human labour, and a means to exploit it, productised LLMs

open.substack.com/pub/agentsde...

Ahead of Germany’s election, with anti-immigrant sentiment rising, they released viral videos showing life without immigrants.

History proves that when a country starts erasing certain people, it doesn’t end well.
🧵

There is such an imbalance between the ability of orgs like Anthropic to create technology and their ability to identify and reason about the true impact or significance of that technology for society. They need, dare I say it, a more diverse set of thinkers and decision makers

😂 you hit the nail on the head here. it’s a huge weakness in using them for more complex “knowledge” tasks as they can’t say “mate - that is just not possibile, try something else”

A Bear Case: My Predictions Regarding AI Progress — LessWrong This isn't really a "timeline", as such – I don't know the timings – but this is my current, fairly optimistic take on where we're heading. …

Smashing the endorse button as fast as I can
www.lesswrong.com/posts/oKAFFv...