DenisM

In my little ai bubble the original blog post was very influential. It does sound immodest, but could also be true.

Register two domains in different jurisdictions, with neither serving its own jurisdiction, and a redirect for stray users.

Not clear where 2x claim in the headline came from, I don’t see it in the linked site.

the runaway inference truck.

Nicely done, thank you.

No sarcasm. In that thread we discussed recursive, interleaved application of code and prompt. No morphing, however.

Thanks for dynamic workflow pointer. I don’t know if I like JavaScript for workflow definitions tho. IMHO sshwarts has the right idea on severely constrained workflow definition language. I plan to look closer into ADK2 workflows as well.

I’m not saying anyone did what you are doing, I’m saying multiple pieces are converging on that, at least in my imagination.

The idea of morphing prompts to code and back is revolutionary.

You may want to read earlier discussions https://news.ycombinator.com/item?id=48881112 And https://news.ycombinator.com/item?id=48051562

I

I tested the language against cold agents

I was procrastinating on that, thanks!

the recursive nature

It’s not clear from you answer if we’re on the same page. Your options are allow the worker agent run wild or constrain it with a rigid guidance. Skillscript can serve dual purpose - constrain the workflow to make it more observable and constrain elemental parts to make them more directional. Apologies if that makes no sense - the AI red shift (expanding universe) makes it very hard to communicate.

1. How a this better than just using any other script language?

2. Inventing a new language complicates large models ability to generate such scrips compared to a well-known language. Did you find it to be a problem? How did you mitigate?

3. The AI is showing. :) I had the similar discussion with ChapGPT and some phrasing is near the same. Not a dig, just a funny observation.

4. Consider the recursive nature of the problem you’re solving - large model updates workflow which you review each time, worker models generate plans and tool calls which you don’t review. A constrained language is useful in both cases to guide the model.

5. This Earlier discussion can provide useful background for why this is needed. You have probably seen it, but the readers will likely appreciate. https://news.ycombinator.com/item?id=48051562

The problem is real, Thank you for taking a stab and sharing your findings.

You seem to inhabit a hero/villaim dichotomy, and a narrative one at that. It’s not a place I wish to revisit.

For A while I was expecting that MCP will dominate, but we seem to be going in the direction of CLI being more prevalent. Can’t wrap my mind around it.

How would you compare that approach to spawning sub agents to operate high-token tools?

It stops being interesting, or even sad, after a while. People get stuck in all kinds of places, mentally. Some get unstuck eventually. It’s only sad if you have come to a counter factual belief that it could have gone better.

I went in the opposite direction - how far can I push myself to see multiple facets of a story? That is a wild ride, and it gets progressively more wild.

Yeah, there’s a Very similar comment, nearly the same words, from another guy.

In short, yes. The effort will amortize for a larger team over longer time period.

https://openai.com/index/harness-engineering/

How do employers perceive such diploma? I would try to find out before spending time or money. Did you?

It’s the diffrence between net present state and the sequence of events the lead to it, aka event sourcing.

There are pros and cons both ways.

NPS is more succinct and devoid of intermediate deviations, thus less likely to throw the agent off target.

ES otoh does contain deviations which might be salient in the future reasoning. If nothing else then to prune approaches that are known to be dead ends, but also fruitful approaches might have use especially when couple with the line of reasoning that lead to their application.

What if a tool produced an error and a retry? Is retry loop now a part of the log?

How did Amazon relocate? I think most employees are still traveling to SLU to their offices.

Can you share more? It’s the first I hear of lora outside research papers. Practical applications would be great to see.

Lora if effective could be a great reason to run local models.

Having gone through encryption export control I came to conclusion that preventing export of technology was never the goal.

By choke-pointing commercial efforts to export cryptographic tech the NSA got a perfect view and control lever of expertise export and transfer.

My hunch is that while deploying An algorithm is easy, deploying a system that is actually secure as a whole end to end requires a lot know how.

Bottom line it’s a bot naive to think that NSA is simultaneously very capable at being an omnipresent spy and is so stupid to not recognize that export controls have obvious holes.

That's not a character flaw. It's a phase.

Oh dear. Must we? Half the article is like this.

did not do much better

Not a joke, actual roundabout.

Under socialism ecology did do much better. Maybe it’s not the ism that’s at play here.

Simpler still, a “minor” bit on the phone, set by parents once. All services must respect the bit in http headers, and app stores should refuse to install certain apps. No need for id check

I imagine that many parent don’t want micromanage their kids apps, this takes care of the problem.

Did you manage to setup a discussion with the agent to reveal such assumptions? Sometimes the shave wrong unstated assumptions when contradicted by evidence, but if we’re taking about a plan for the future the evidence is thin.

In some ways they are giving you a spec and UI mocks to implement.

with the hope of it giving you the right answer with the right prompt.

Consider that our ability to evaluate quality of the output is falling further behind our ability to produce it. The “right answer” is not the most likely outcome.

Notably the feed metrics and traces into the agent. I never thought of doing that.

Fantastic job!

Can you share what type of project that was? On the spectrum from a database engine to cat picture sharing web site (very high demand for correctness vs very lax).