LP

lol Block/Square stock is up like 20% on laying off 40% of their headcount while growing and profitable. The Faustian bargain that big tech made with society was by creating a nuevoaristocracy of laptop-class landed gentry, their founders and shareholders were allowed to become wealthy beyond all imagination with little government intervention past slaps on the wrist. Google could cut 90% of the headcount and generate 90% of the revenue. But when this deal is broken, the displaced Baron’s will demand their own Magna Carta - tech leaders are too politically myopic to see this and so the more likely outcome is one of revolution. A rising of modern sans-culottes currently kept busy by the bread and circus’ of Doordash and TikTok but rallied to action by these displaced elites. AI may be the excuse but this has always been an option. Salesforce employs 75,000 people. Including Matthew McConnaughey on over $1,000,000.00 a year. Most of this headcount can hit the road - Covid proved many jobs are ritualistic button pressing exercises in a circular loop of keep up appearances financed by a need for elite docility. AI’s biggest disruption may be giving people the ‘excuse’ to finally break this pact. But I assure you the Maximilien Robespierre of today is not a puffed up perfumed Frenchmun. He’s a forty year old Stanford educated father of 4 who contributed to the Review about to lose his director role in Big Tech and enter the worst job market since the financial crisis. He will likely be equally as unforgiving if executives are not careful about how they carry themselves and their decisions in this coming shift.

Sub agents and agent modes are crap. This is getting so dumb. I asked claude to implement a feature, so first it switches to plan mode, then it calls the explore agent, then it does some more random thinking and calls, then it calls the plan agent to create a plan. This is AI theater, none of this helped and so many calls were duplicated. This could have just been one linear thread. And then it by default want to "clear context." You know what the first thing it does after you clear the context!? It re-reads all the same files it just did. THIS IS SO STUPID AND WASTEFUL. This is what happens when engineers think they are smart. You don't need multiple agents, your just making it worse.

Poor man's continuous learning: How to make agents better without fine-tuning or retraining. Over the last few months, I've been using a simple pattern that's made my agents noticeably more reliable and useful. It's also been the most fun I've had building in a while.