farez 🇵🇸

KL folks should get in on this too

We are entering an era where human-maxxing will be rewarded. Given infinite AI appslop marketed by AI adslop, the way to win is: - be human. your marketing is you and a camera. - app pain point is something personal to you - landing page is not fucking purple

@yongfook That Jimmy Savile colour palette really needs to die

When we all have agents running around doing most of our tasks for us, even AI summaries will stop being used by humans. For SaaS founders, I think at the moment we're somewhere in between "we know what's going to stop happening" and "we have an educated guess where things are heading".

And yes I'm interested. I serve service businesses and freelancers. My product, popinvoice.com, automates admin work when it comes to generating estimates, invoicing, collecting and chasing up payments. My customers are dynamic business owners who care about saving time and money, and care about the quality of the services they provide. Their goals are to continue growing their businesses while keeping a small tight team to run them. If you have a product that can serve them, and if you think my product can serve your customers, then lets chat!

I think one of the best things AI product founders can do right now is partner with each other. Cross promote each other's products. Upsell other fellow founders' solutions to your existing customer base. 4-5 founders bringing in their combined customer base into their own ecosystem of related solutions.

@0xishans Signed up for Friday’s event

London AI Events this week Catch me in some of these:) Leave a comment if I missed any👀 Wednesday: Vercel Meetup vercel.com/go/rewriting-t… Thursday: Hugging Face London #4 luma.com/aeup01om Friday Day Time: Co-Working Fridays : London Chapter luma.com/london-mar-06 Friday Evening: OpenClaw Meetup with Peter Steinberger hosted by OpenAI & Sequoia Capital luma.com/ierc56lp?tk=Va… Saturday: UK AI Agent Hack Demo Day luma.com/54b06cos

Spec + Data Not a fully formed thought, and apologies for the brain fart, but... This post was inspired by an article on @latentspacepod how, first, human coding "died". Then now we're going to see the demise of the "pull request". If you're not familiar, a pull request, or "PR", is a request to change some code in software, by introducing some new code that you have written yourself. So someone has to review that code, make sure it's all written "properly" and then approve or reject it. Code can now be generated 1000x faster by AI. And AI generated code still needs to be merged into the main code. And if every chunk of AI generated code needs a pull request, that's thousands of pull requests to review by humans. We then become the bottleneck. Then why not get AI to do the review? It can do it much faster than humans. The @latentspacepod article asks why have pull-requests at all? It was AI generating code. Why have the same AI look at it again? Personally, I disagree, for now anyway. I think we're in a transition period where humans still need some control. And while we don't write code anymore, and I do believe PRs should be reviewed by AI, I think humans should still manage the PR agents - for example defining rules for the review agents to follow, checking that they did follow the rules, etc. But that's transitional, and besides my point. What I do agree with in article is where humans will be needed. And where we're needed is further up the software development chain, into the realm of requirements and testing. Code is not created in a vacuum. It is created to automate work. It is created to fulfill a specific requirement. And this requirement will still be created by humans. And if we can be clear about what we need - the requirements - then we can also run automated tests to check that the code meets our requirements. If code is cheap, and instant, and humans stop become the bottleneck to deploying code, what does this mean for code ownership? It means it doesn't make sense to "own" code anymore. So what's worth owning? Spec and Data. How your organisation runs, all its rules, all its goals. These make up your software specification (spec). All the information your org collects, all the information it needs to run. That's your data. If you have a clear, tight, spec. And a database that holds all your data. Then you can reproduce all the software that you need instantly. An AI can read your spec. Then create automated tests against that spec. Generate all the code needed. Connect to your data. Test the code against the spec. Verify that it is all correct. Deploy your code. And hand you the keys (login) so you can drive it straight away. This has several further implications: - Should an org's investment now turn to speccing and testing? - Should we work on a more standard way of defining spec and data schemas? - What does this mean for software development agencies, if producing and reviewing code is not needed anymore? - When an enterprise buys from Google and Microsoft, what are they really buying, if it's not software anymore? Definitely more questions than answers. And things will continue to change fast for the software development industry. And for organisations - if you already create and document great spec, and have automated tests in place, and have good data policies, then you're in a great place.

@Dotnetster Ooh need to look it up.

The #diff Surprise Film this year was ... "Dead Man's Wire". Best surprise film since "Get Out". More of this and less of movies like "Dragged Across Concrete" please.

@0xishans @steipete @openclaw Damn it. How did I miss this!

Wow @steipete & @openclaw in London London isn’t that bad

@elonmusk Civilised discourse isn’t your strong point Elon

Anthropic hates Western Civilization

@donaldnzy Awesome! Definitely will try it out for OpenClaw.

@farez Yeah it is my openclaw main driver now. Good enough for most daily operational stuffs. Can't really complain for the cost!

This daily weather update from my OpenClaw costs $0.13 per task. That's a total of $47.45 per year. Too expensive for such a simple daily task. For simple tasks, a dedicated app is probably cheaper.

@AnthropicAI Hey @elonmusk, what would you have done?

A statement from Anthropic CEO, Dario Amodei, on our discussions with the Department of War. anthropic.com/news/statement…

If I was a better engineer or entrepreneur, I'd achieve more with them. AI makes you realise where your knowledge and experience gaps are. E.g. If I was better at creating tests, I'd be able to instruct Claude Code to do it properly. E.g. If I was better at marketing, I'd be able to instruct OpenClaw to automate what's already working.

@farez @claudeai @openclaw What do you mean?

Claude Code exposes my weaknesses as a software engineer. OpenClaw exposes my weaknesses as an entrepreneur. @claudeai @openclaw

@donaldnzy Heard a lot about Kimi. Are you using that regularly?

@farez Use a cheaper model like Kimi K2.5 which is very capable for most general stuff and smart "enough" for coding too.

I wanted Claude Code to work on a large development task before I went to bed and experimented with the following: 1. Ran Claude with --dangerously-skip-permissions. If you're not connecting to live systems and working in a new git branch, this isn't too big a deal and a great unlock. 2. Gave it all the task broken down into steps/phases. No surprises this is a good way to prompt. 3. Gave it autonomy by adding this to the prompt: "If there is any uncertainty about the details of the task above, make an assumption and make a decision that makes sense, and note it down for me later. At the end, summarise everything you did. And then include any assumptions and decisions you made so I can check them." I think that last prompt helped give it more autonomy but still allows me to, later, review those things that would have required my input during the work. This morning it completed the work successfully, and in its summary (see screenshot) showed me the decisions it had to make without my presence. Super useful.

@karpathy I’m assuming you’re using OpenClaw for building. What’s your setup? I’m mostly just a “one Claude coder” but would love to see examples of “claw orchestration” for building faster.

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow. Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes. As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now. It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

@paulmichaeldev Maybe this thread is costing us more brain tokes that it should lol

@farez I KNEW you'd focus on that! 😂 My actual point was on the cost of a "simple" task.