@zephyr_z9 June 22nd, Anthropic sent an email to subscribers about an update to their Privacy Policy, in it:
"2. Verification data. As part of our measures to keep our services safe and secure we may ask you to verify your age or identity, and we've described what we collect and how."
It’s a bit telling that I listened to @davidsenra‘s @FoundersPodcast daily for a year straight then started a company.
Respect where respect is due @davidsenra, excellent work.
If you can scalably understand the model queries then you can A/B test response preferences against user subsets and effectively tailor responses to the user. Language is thorough and composable so you can get a very fine tuned response profile.
@AviFelman They're not "space" assets, just as we wouldn't say "Earth" assets. SpaceX just has technology that it deploys in space to serve industries on Earth. The assets are tied to the industries they can serve, not space.
@VictorTaelin It's unfathomable for a company to give away their golden goose which means gov/s would have to pay the company 100 billion+ or commandeer the tech. The gov buying the tech is the only one that doesn't have immense negative impacts for the industry, but would never be approved.
There are 2 narrowing windows of arbitrage:
- The expertise to effectively harness AI to create quality code quickly
- The lack of real adoption/integration of AI by incumbents
@zachtratar Using Gemini 3.1 something similar to this worked for me yesterday. It had failed multiple attempts at debugging and then I gave it a similar prompt and it fixed the issue.
@cremieuxrecueil For VCs, their returns really matter when pitching their investors. I would suspect a VC publicly showing loyalty to a fraudulent company is due to the VC placing too large of a bet in that company and being stuck in the dilemma between good morals and future job prospects.
@VictorTaelin It would be nice if benchmarks had an upper and lower threshold that showed the model's performance when the prompt is lazy/bad vs. when the prompt is designed to drive the best performance.
Overtime the difference should be negligible but rn there's material variance.
My brother operates a few Airbnbs, he recently implemented an AI auto-responder that will escalate to him when necessary. He says, his time requirements have gone down 50%+ and customer reviews have been steady.
It's working.
@ramahluwalia The latter is much less about the bankruptcy attorneys and more about bankruptcy law, the attorneys will be mandated to retain control of the assets in liquid form ASAP at whatever price that brings. Their wiggle room for speculation is weeks to months not years.
@GeminiApp would really love if your interface had a scrollable sidebar area for the chat outline (like docs), except the prompts are the default headers/outline items... I'm getting really tired of having to scroll, scroll, scroll up to find a reference.
@cremieuxrecueil This can be misleading because people want to infer preference from the chart. But, the chart doesn’t normalize for the population of each ethnicity.
For example, if Indian Males married Indian 63% of the time and Indian females are 63% of the population then that is expected.
@duolingo has an unruly amount of annoying click-throughs after a lesson. Here are 7 that I had to click through after a lesson.
It is easier to close/clear the app and then re-open instead of clicking through them.
If the company executes to the limit, the outcome is inevitable. Any government action will send shocks to the market. The crucial thing is the timing and force of government action. Miscalculation in either dimension can lead to persistent negative impacts to the economy.
In a capitalist society, a company’s power is limited by execution or government intervention. If a company continues to execute it will outpace its peers and grow in power. If the company does this for long enough, the government is forced to betray the free market.