ChicagoDave

I think domain specific harnesses are already surpassing generic harnesses. I also think software development is its own domain.

My harness is a Claude Code plugin with its own brainstorming, adr, and planning skills with associated review and interview skills. Behavioral testing related to acceptance criteria is built in. Everything in my harness is gated to prevent ratholes.

I recently inflated a docker container to execute a set of work with Claude in unsafe mode and immediately saw problems with everything it was doing…and then I realized I had not installed my harness.

Running Claude without an engineering harness is like driving a car without brakes or a steering wheel.

I’m still amused at how many people will lick the boots of any oligarch/startup toad.

I played around with Neo4j and only ever found two domains where it excelled.

- custom app security

- social media

I also think cypher is a brilliant way to query a graph.

Does anyone at OpenAI know how to run a business?

Shipped version 2.2 of my Interactive Fiction platform Sharpee with Phrase Algebra.

Designing a new DSL (Chord) that compiles to Sharpee (Typescript).

https://sharpee.plover.net/

A lot of people died from Covid and if not for mRNA technology and extraordinary caregivers a lot more deaths would have occurred. That’s hype where it truly belongs. Don’t mix AI hype with Covid conspiracy theories.

They are also the most expensive operational data stores and the younglings know almost no SQL or set theory.

Schema management is the single worst part of deployments.

If I build on DynamoDB vs RDS I can save 10x.

I extracted my books from my kindle over a year ago. Never buying books from Amazon again. EPUB on iOS works great.

This implies nosql data stores are not ACID capable. Mongo is fully capable and DynamoDB is mostly capable.

I would challenge you to look at event driven architectures, CQRS, event sourcing, and how to implement and leverage read models.

It will expand your architecture toolkit.

Yes and for crud systems relational is fine because you're unlikely to over-complicated your architecture. But when a system starts talking to other systems and its bounded contexts become complex, alternate solutions should be sought.

The problem with "schema change", and I did this for decades, is that it's always a massive blocker. In some companies the data architects had to approve and implement schema changes. You could wait days for that. NoSQL allows you to modify the document surface in mostly non-breaking change ways OR it's easier to version your APIs to handle different document versions.

Simple CRUD: Any data store is fine. Complex multiple bounded contexts: Choose the appropriate data store for each bounded context accordingly.

My point was no one should be reaching for a relational database or starting with an ERD to build a system. Document behaviors. Model the system. Let the system decide what data storage it requires.

Sorry I missed it. I was expecting an As clause but conflating VB is probably not what is wanted.

This is very cool, though I'd want to see typed FUNCTION blocks.

Data consistency was solved in Mongo and DynamoDB years ago. CQRS is a better pattern. Read Models out of analytics (relational) data stores are better for dashboards. I stopped being "SQL First" ten years ago and never looked back. Saved clients time, money, and improved maintenance and eased feature additions.

ORMs taught me that relational databases are an operational anti-pattern.

NoSQL for operational data storage is more efficient and cost effective.

ORMs were a regression test that exposed unnecessary complexity.

I bought the 16” M5 Pro 48gb 1tb black at Best Buy with the $150 membership savings.

Once I move from the mini to the MacBook it looks like I can get $800 for the mini. I got 0% financing too.

I was inspired by this. I made a Claude skill to take my resume and compare it to any job description to point out viability and gaps. Pretty cool skill. I'll post it somewhere.

I was literally ordering an M5 MacBook Pro tomorrow. The total is $900 more now. Might have to hold off and just live with my M4 Mini.

I've not seen anything from the base models that replaces my engineering harness (workflow). There's a significant gap between what a generic LLM does and the domain I work in (software construction for complex applications).

Understandable. Look at the possible directions:

- GenAI becomes a foundational requirement for tech and non tech sectors. If you’ve refused to engage, you’ve self-selected out of any of those sectors.

- GenAI usage shifts down to just the tech sector, but in an integrated fashion where current engineering practices are still desired. Everyone survives, but pay scales are adjusted down by a not-insignificant amount.

- GenAI bubbles badly, OpenAI and Anthropic merge with Google/Microsoft/Oracle/IBM/???. Tokens become extremely expensive and no one is leaning into agentic integration. Everyone thrives.

There are a number of reasons I’d site for the current job market tightness:

- political: there’s an enormous amount of uncertainty here. All businesses make plans and uncertainty puts them all on pause.

- economic: related to political, but we’re teetering on a very bad recession. Watch where national oil reserves go.

- AI: I throw this in with every new technology that comes out. There is always a period of chaos before normalization. We’re still in the chaos phase.

- Business Pain: Right now I don’t see any sector that’s in pain. Inflation has hurt consumers, but we’re still spending. When consumers lock it down, that pain comes back and job market shifts with it.

I have no solution other than figuring out a way to do your own thing. There’s no better time to be a founder.

I have two sessions going. It’s my fault. Sorry guys.

Sadly, the API isn’t ready yet so tooling has to wait until 7.1.

Well I was riffing off the xkcd free speech comic.

https://xkcd.com/1357/

This sounds like BlueSky without the user-directed self and grouped moderation tools.

That’s not censorship. It’s showing you the door if you’re a jerk.

And it sounds awful.

I guess I just assume a good architect would enforce authorization with an identity server and never put anything important on the client.

This is the most undeserved startup buyout in history. Literally anyone with a fork of VS Code could replicate Cursor in a very short period of time.

OAuth2 has been the web security standard for ten years. I don’t recall any major security announcements about it. The OP is confused.

The OPs domain/subject matter expertise is the part that should elevate their career. Understanding how large applications are constructed should also remain a pillar.

The coding and debugging part will be GenAI and possibly guardrails (harness engineering) tuned specifically for fintech, which they are also well-suited to implement.

The second I realized it removed nearly all blockers as a bootstrapped technical startup founder.

Claude wiped out the need for web and mobile development resources. I bought a Mac-Mini and had iOS apps up and running in days.

If it’s a woman slapping them off a pervert and accidentally stepping on them, 100% walks no charges.