dreamingof.eth

"When all this works, the result is astounding. Small teams of highly competent managers-of-one can progress at an unbelievable pace. And there's nothing better than shipping quickly alongside peers you respect highly." world.hey.com/dhh/we-once-mo…

> "Ruby isn't high performance enough for us" > Their website downloads 7mb of JavaScript to display a blog post, makes 3 network requests in a row to hydrate data, re-layouts twice

Quantum Entanglement in Your Brain Is What Generates Consciousness, Radical Study Suggests apple.news/Aa_YBGChLSYGPM…

After trying Next.js for my own company, I decided to fully go back to Ruby on Rails. I'm amazed by the good people from the Rails community and believe we'll have a bright future using it. Thank you, @dhh and the Rails team. blog.wildcat.io/2024/08/i-m-ba…

These types of symptoms will become more and more common

🥲🥲

@kiranadimatyam I just experienced the uncanny valley.

Robort, the movie! Watch it when you have 10 mins time. youtube.com/watch?v=N_Nvr4…

Solid Queue has finally conquered all of HEY. Byebye to Resque! It served us well over the years, but it was due time for a clean-sheet implementation that could compress all of the accumulated complexity. Tens of millions of daily HEY jobs now run off SQ exclusively! 🤘

This Robin Williams thread will brighten up your day 🧵 1. His improvised scene in Good Will Hunting made the cameraman laugh so hard the camera started shaking. Matt Damon's reaction is authentic.

Incredible! AI brings celebrity then and now photo to life

The 12 things getting in the way of your execution: 1. Tons of meetings The counter-intuitive thing about great execution: it generates heaps of urgent, last-minute tasks. You must be available to track down different strains of items to unblock your engineering & design partners. If you're booked with meetings, you're slower. 2. Staying away from the details The heart of great product execution is the details. · It starts with data on user flows before your feature · Then during the feature, it's about usability testing · And once it launches, it's about analysis If you stay away, you learn less. 3. Under-resourcing design One of the most common drivers of poor execution is a low ratio of designers: engineers. Designers have to move extremely fast. Instead of spending time mapping out all the options and flows, they make quick decisions. And due to this, execution suffers. 4. Avoiding usability testing In the rush to get things out of the door, it's tempting to avoid putting the product in front of users. But getting designs in front of users is cheaper than ever. Skipping this step increases the chances of the feature failing out in the wild. 5. Skipping UAT Too many PMs are keen to let QA or engineers lead the building of a feature. But having a user acceptance test is a great way to use and test a feature before customers do. It keeps you close to the details of the user experience. 6. Writing vague PRDs Another one of the most common drivers of bad execution is poorly written PRDs. Pushing speclets without defined user flows and messages to developers results in them making decisions on the margin. Or missing the corner cases altogether. It drives detail errors. 7. Leaving analytics to the analysts There's simply too many product execution details for analysts to do everything. The really "deep" execution PMs query the data to fill in the gaps. It's especially helpful to increase speed to event spec'ing for developers. 8. Forever planning If you can't spend time away from planning cycles, you won't ever be able to dive into product feature details. Regularizing planning cycles and avoiding "forever" planning allows you the time and space to crush the details. It's critical to execution. 9. Being allergic to asking about the status It feels unproductive to ask "how is building X going?" But it's actually critical for PMs. Having regular dialogue with designers, developers, & analysts will unlock execution wins. 10. Building for everyone Having specific users in mind for a feature helps you identify the different ways they can deviate from the happy path. Putting on those user hats - being specific - helps you get every flow & detail right. Otherwise, it's easy to miss a key use case. 11. Having a bad strategy You can have the best execution in the world. If you strategy is flawed, it's not going to go anywhere. If your strategy is the problem, great execution means spending the time to improve there. Define your user problem well with data. 12. Not repeating yourself Execution is a team sport. To really help the team internalize and realize a strategy, you have to be willing to repeat yourself. You must be able to say a thing different ways to drive home the key objectives. It's the essence of PM communication.

Gymnastics is the Turing test of video generation models

⚡ The best GEN-3 prompts so far. Runway released its new text-to-video AI Generator 48 hours ago. I've spent hours testing the most advanced and craziest prompts to get the ultimate results so you don’t have to. Here are ALL of them (+45 prompts) 🧵👇

“The Missing Link in Corporate Decarbonization: Effective Tools for Setting Emissions Targets” by Jason Denner jasondenner.medium.com/the-missing-li…

The Supreme Court's Chevron ruling may be most impactful things to happen to startups in a long time, in ways that people don't realize. A thread:

every single day for four years I have thought about Nazi jurist Carl Schmitt gloating that you can dismantle a democracy right in front of a liberal's face piece by piece and all they'll do is convene breakout sessions and committee meetings til the boots are in the halls

Does AI code prefer tabs or spaces? Camels or Snakes? Singles or Doubles? Short variable names? Does it even care about being consistent in any of these? If we ask AI to explain the code it has written, will we even care? I don’t think so.

@strzibnyj @dvassallo Me too!! Looks cool. I am ready to get inspired.

I finally joined SmallBets by @dvassallo and thus I can also finally experience Campfire from 37signals 😅🙏

This is an exciting discovery! The collective intelligence of multiple open-source LLMs outperforms GPT-4o. "Together Mixture-of-Agents" leverages diverse models to achieve 65.1% on AlpacaEval 2.0, surpassing GPT-4o's 57.5%. together.ai/blog/together-…

@hiakkio @rezkhere I am curious please send.

I make more than $15k/month with my AI automation agency. No degree or coding required. I made a complete guide on how to grow your AI automation agency from 0 to $15k/month within 90 days. To get this, comment 'send' and I will DM it to you (must be following me)