Kerry Jones

I am going to Bootstrapping / Vibing Founders Meetup. Join me! luma.com/23rpub7i via @LumaHQ

The next Bootstrapping / Vibing Founders Meetup is booked in for 6th May in London. Come and meet other founders building their businesses using AI. luma.com/ca3v45kx via @LumaHQ

@ryancarson And since you started using Symphony and thinking about this post, now paperclip has been opensourced? :)

Be careful about reading/believing articles on X. Including *my* articles. As soon as anyone posts, their thinking/work immediately evolves and iterates, and the article is instantly out of date. Example: I need to post a brand new version of the Code Factory article (1.8m views) because I'm now using OpenAI's Symphony spec. Everything changes so freaking fast.

I am going to Bootstrapping / Vibing Founders Meetup in London on 24th Feb. Join me! luma.com/23rpub7i

I am going to Vibing Founders Christmas Drinks. If you're a solo founder with no Christmas party to go to and will be in London on 9th Dec, Join me! luma.com/9a4bqs0u #Startups #solopreneur #Entrepreneur #london #buildinpublic

@HamelHusain 😂

Dear anyone that said RAG is dead, Eat it.

Introducing Cursor 2.0. Our first coding model and the best way to code with agents.

@sqs @jkeesh @nikitabier @Amp @AmpCode @XHandles Booooo

It’s a wild time to be building. AI makes it easier than ever to start something new. Come meet other bootstrapping founders figuring it out, experimenting, learning, and swapping stories in person. luma.com/8dgpgz6a

If that’s you — come join us this Wednesday, Oct 22, in London. All bootstrapping founders welcome. 👉 luma.com/8dgpgz6a #bootstrapping #founders #londonstartups #solofounders #aicofounder #buildinpublic #meetup

That’s why we started the London Bootstrapping Founders Meetup — a space for people actually doing it. Founders figuring it out, experimenting, learning, and swapping stories in person.

Over the past year, I’ve noticed something interesting in London’s startup scene. More and more people are quietly building on their own, using AI to move fast, testing ideas, and chasing revenue without funding or a co-founder.

@ryancarson I hope the agents you’re building are designed to build their own new agents as well then. 😂

@ryancarson It’s fascinating. Out of interest, are you doing this all in Amp and deploying your own code to do things like customer support? Or using other off the shelf products.

Adding tools to an agent is so fun. It's amazing to start with a very competent, hard-working, eager-to-help, very knowledgable employee. Then you start putting them through training courses and giving them extra functionality. Then you get to see them use those tools in the wild to help your customers. What a time to be alive.

@aiwithjainam Send

Fuck it. I'm sharing the 10 mega prompts that helped us automate all our work. • Strategy • Research • Marketing • Planning • Delegation Comment "Send" and I’ll DM you the file. (Follow me to receive it)

@ttorres Hearing that from a few different people today.

Is it just me or did Claude Code get way worse this week?

In era of pretraining, what mattered was internet text. You'd primarily want a large, diverse, high quality collection of internet documents to learn from. In era of supervised finetuning, it was conversations. Contract workers are hired to create answers for questions, a bit like what you'd see on Stack Overflow / Quora, or etc., but geared towards LLM use cases. Neither of the two above are going away (imo), but in this era of reinforcement learning, it is now environments. Unlike the above, they give the LLM an opportunity to actually interact - take actions, see outcomes, etc. This means you can hope to do a lot better than statistical expert imitation. And they can be used both for model training and evaluation. But just like before, the core problem now is needing a large, diverse, high quality set of environments, as exercises for the LLM to practice against. In some ways, I'm reminded of OpenAI's very first project (gym), which was exactly a framework hoping to build a large collection of environments in the same schema, but this was way before LLMs. So the environments were simple academic control tasks of the time, like cartpole, ATARI, etc. The @PrimeIntellect environments hub (and the `verifiers` repo on GitHub) builds the modernized version specifically targeting LLMs, and it's a great effort/idea. I pitched that someone build something like it earlier this year: x.com/karpathy/statu… Environments have the property that once the skeleton of the framework is in place, in principle the community / industry can parallelize across many different domains, which is exciting. Final thought - personally and long-term, I am bullish on environments and agentic interactions but I am bearish on reinforcement learning specifically. I think that reward functions are super sus, and I think humans don't use RL to learn (maybe they do for some motor tasks etc, but not intellectual problem solving tasks). Humans use different learning paradigms that are significantly more powerful and sample efficient and that haven't been properly invented and scaled yet, though early sketches and ideas exist (as just one example, the idea of "system prompt learning", moving the update to tokens/contexts not weights and optionally distilling to weights as a separate process a bit like sleep does).