rogerkirkness

Depends on whether you like have parasocial relationships with machines or not. It will say what is necessary to get you to keep using it on a regular basis.

the weights are the bones

Still, in the end, open-weight models deter further AI capex.

Good.

Around 2002 there was way more internet bandwidth and infrastructure available than utilized, eventually (not immediately) leading to business models like Netflix current one. I do think a lot about what the AI analogy will be.

Apple would punish him severely unless they cleared it in advance, it might be to their advantage for some reason (negotiating with Google for Gemma rights? idk).

Claude Code is bad UX that tries to rip work out of people's hands instead of actually being aligned and figuring out how to increase their differentiated expertise. It's about renting your future from Anthropic rather than building it or embodying more knowledge yourself.

I don't think so. We aren't hiring right now, but definitely add us :) Feel free to email the email on my profile.

32 hours full comp.

Our startup implemented four day work week without changing comp, and submitted various data as part of a recent Boston College study. Both wellbeing and output increased in statistically significant way with a team of only 15. Hard to avoid confounding with agents adoption, but overall highly recommend it.

Canada has tens of trillions of dollars in natural resources it could choose to monetize at any time.

It makes no sense to do broad layoffs during sweeping technical change that affects you, your customers and your competitors all at once. Makes way more sense to cling to your best people and re-train them in order to preserve your soul and institutional knowledge. I think deep layoffs + regret as opposed to re-training is actually uniquely American culturally and not a thing elsewhere.

Shopify is a pivot from selling snowboards online in Canada

Going double or nothing using a business that invented reusable rockets and makes $8B a year in profit as the bet is overwhelming to think about.

You can see the business cycle each time there's a giant spike in the amount of debt carried by institutions that isn't paid down before the next one.

It's really tragic given how awesome SpaceX is as a body of work and financially to merge it with such a terrible disaster in X.

Yes. There's already rewrites 'inspired' by the now public code as well.

I've met several of the people on the first few pages on the watch list, and they are among the sketchiest Silicon Valley people I've met. The criteria are plausible.

Maybe I'm not smart enough to be a customer, but the testimonials is the most contemptful thing I've ever read.

If you're so smart, why can't you explain why this is useful with smaller words?

@dang seems like AI? Would just ban

I'm glad this story is spreading. On one hand, super cool. On the other hand, this is probably the #1 most risky aspect of LLMs wide distribution. Someone working on proteins with good intentions creating the next Covid-19 or worse.

I assume it means changing governance policies while letting them continue to make their own decisions within that framework.

Yeah one is historical revenue, and one is projected revenue, and it doesn't sound like the definitions are mutually exclusive or incompatible.

In terms of just the US part of what's going on, this sounds very accurate.

Fast takeoff.

I own a Tesla with FSD I bought in 2021 so I feel that. It clearly is far worse than Waymo still.

Not a fan of Elon but he said his job is 'To turn impossible engineering projects into late ones' and I thought that was pretty accurate.

The best example I can offer is that when given a marketing goal, a subagent recommended hacking the point-of-sale systems of the customers to force our ads to show up where previously there would have been native network served ads. To do that, assuming we accepted its recommendation, would be illegal. My email is on my profile.

We're a startup working on aligning goals and decisions and agentic AI. We stopped experimenting with decision support agents, because when you get into multiple layers of agents and subagents, the subagents would do incredibly unethical, illegal or misguided things in service of the goal of the original agent. It would use the full force of reasoning ability it had to obscure this from the user.

In a sense, it was not possible to align the agent to a human goal, and therefore not possible to build a decision support agent we felt good about commercializing. The architecture we experimented with ended up being how Grok works, and the mixed feedback it gets (both the power of it and the remarkable secret immorality of it) I think are expected outcomes.

I think it will be really powerful once we figure out how to align AI to human goals in support of decisions, for people, businesses, governments, etc. but LLMs are far from being able to do this inherently and when you string them together in an agentic loop, even less so. There is a huge difference between 'Write this code for me and I can immediately review it' and 'Here is the outcome I want, help me realize this in the world'. The latter is not tractable with current technology architecture regardless of LLM reasoning power.

I think there's an argument where if Claude had the knowledge map of your personal one liners and a tool for using them, it would often do the right thing in those cases. But it's definitely not as able to compress all the entropy of 'what can go wrong' operations wise as it is when composing code yet.

This is written by AI.