Ryan Colbert @rmcolber

@ArthurMacwaters @elonmusk @grok is this post from @ArthurMacwaters accurate? Which countries have mail in voting provisions. Compare them to the US. Build a table of key features and restrictions. List all counties showing their status for those features and restrictions. Show sources.

What’s interesting about AI agents is the unit of a task is just going to keep on growing in size over time. In code, in the past 2 years, we’ve gone from autocompleting a couple lines of code in a second, to writing hundreds or thousands of lines of code in minutes, to writing tens of thousands of lines of code over hours. The same trend is happening right now in most other fields of knowledge work. Outside the space some may wonder “why isn’t AI getting cheaper or faster?” Yet, everyone in the industry knows that we’re now able to throw more compute at a problem to solve the next harder one. As soon as we think we’ve hit an optimal point of quality, price, and speed, the use case potential just expands as we get more model breakthroughs. The last problem we could solve looks trivial by comparison, and we’re onto the next complex task. This dynamic will likely keep happening for the foreseeable future given we’re nowhere near the plateau of capability, so the task units will just keep growing and growing.

This aged well.

Once we get arbitrarily long context windows that have perfect accuracy, AI will be able to do literally anything. This is probably the most important gating factor for AI right now.

I'm in a judgy mood. So here's some things I'm gonna do with my AI Personal Assistants since that is what OpenAI and Google are trying to bring to us. 1. Does it do the basics? Like when you ask Siri to take you back to your car, it doesn't work, but if you ask it to take you back to your parked car, it does. If you ask it 1,000 questions how many does it answer accurately? In my perfect world it would have no mistakes. 2. Is it true multi-modal? Most of the demos we've seen are just with camera and microphone. What about the barometer inside your phone? Or the IMU motion sensors? Or the heart rate sensor inside your watch? 3. Is it true multi model? Can you ask it to ask it to ask your question, like "please write me a college-level essay on the formation of the United States but please ask Google Gemini, Meta Llama, OpenAI ChatGPT, Anthropic Claude, and please tell me where these differ?" 4. Is it a full citizen of the hardware ecosystem you already own? Can you talk to Apple Music, for instance, and ask it to play your Hip-Hop playlist on any speaker or headphone? Can you control all your devices? How deeply? 5. Can it code? With a big context window (Google just announced two million token window). Like drop your whole code folder into Siri and ask it to add features, debug, etc. 6. Does it remember? Can it look through your old photos and learn things about you, including figuring out your favorite restaurant and what you usually eat there. My wife takes pictures of all her food, for instance. Can it have a conversation with you about that "hey, looks like you like pepperoni pizza, do you like any other pizza?" 7. Can you run it long term? Like leave your phone on a counter during a fun discussion with my family, half of which speak Farsi. Without killing your phone fast? 8. How does it treat your private data? Like if you start talking to it about your mental illnesses, where does that all go? To some server? Do you have control of the fact that it figured out your mental illnesses? (ChatGPT is used by my psychiatrist already to listen to our therapy sessions and it picks up everything, so AIs can pick up lots of weird things you might not want others, or, computers, to accidentally reveal. 9. Can it talk to you about real-time news? Like if a neighbor's house catches on fire, something that happened last week in my neighborhood, how soon can we talk to Siri about it? Can we share pictures/video with others to enhance the news? 10. Can you control AI agents that will do research for you, or build a workflow for you? "Hey Assistant, can you send an agent over to the stock market and look for companies that have a P/E ratio of less than $20?" 11. Can you get one to start up a committee easily to discuss something. "Can you build me a committee that would be good to discuss where to eat in Las Vegas?" (It should give you a committee of food critics, chefs, and restaurant owners). 12. Do you have prompt control of how it appears. "Can you switch your form to Spongebob?" 13. Can I share my voice with your assistant so that it can read my emails to you in my own voice? 14. Can you control every application you already use, whether it's from Autodesk, Adobe, X, or, really any company? Can it instantly get smart about any new thing, like can it figure out how to export a 3D model from Autodesk and incorporate that in the right spot in a spreadsheet or slidedeck you are building? 15. If you have a Spatial Computer can you have it take on a real form (like does it look like a real human being walking around, or a real Coke Can that's talking to you on the counter)? Can you pick it up, or, if human form, high five it, while talking to it? Can it teach you to line dance? 16. If you have a Spatial Computer can it show you how to do something, like fix your washing machine that has a clogged pump, on top of the machine itself? 17. Is it really smart about your context? Like are you in a grocery store looking for some cold medicine? "It looks like you are in a grocery store looking for medicine because you are sick, can I help?" Which should give very different answers than if you are in a classroom learning to make aspirin. This goes beyond just looking at GPS, but looking at other signals about what you might be doing. 18. Does it understand my son and his classmates, some of which have a rough time speaking. My son is hard to understand by humans, for instance, so computers are frustrating for him to speak to. Yes, he can type, but we can train computers to understand people who have speech impediments or accents much better. 19. Can it make calls, get restaurant or hotel reservations, or listen on hold until some human picks up and then talks with them and only gets you involved if there's something you need to decide on? 20. Can it do searches for videos that have to do with your context (where you are, and what are you trying to do). If you are standing at Yosemite National Park on Glacier Point, which looks down at the valley, will you find the video I did with Ansel Adams' son there and let you know about it as something that might enhance your visit? Can you see a park ranger who left a class about what you are seeing there? Or are you gonna get all sorts of regular videos there of people's vacations which aren't nearly as interesting? 21. Can it help you do all tasks in the kitchen, including automatic shopping? "Hey, I noticed you are running low on milk, we'll add that to your next Amazon grocery run." 22. Can it start with a list you make, like a music playlist, or a list here on X, and expand it. "Nice classical list that has only Dolby Atmos music on it, here are 10 more tracks that fit into this list." 23. Can I really talk to it about medical and mental health concerns, even ones that will get me to go see a doctor? Can I buy sensors to expand this? 24. Does it really help me at work? Working with legal docs, finding collaborators, building not just marketing plans, but entire projects. 25. Can it keep track of tasks, particularly on a project. Like if you are planning a vacation for your family you will need to do things like make sure your passports are up to date, buy flights, buy hotels, get rental car, etc. As you do each does it keep track and does it know when you actually want to work on these tasks. "There are still some things to plan for your vacation, when would you like to work on those?" 26. Is there any AI Safety stuff (AKA "woke AI") that keeps me from doing what I want to do? 27. How well does it work in noisy environments, like when I go to a music festival and am standing next to the subwoofers, can I still talk with my assistant? 38. How well does it work in conditions where there isn't any connectivity? For instance, in Sidges, Spain, my wife and I were in a restaurant that was underground and we couldn't use our phones to do translations of the menus to English. 29. How is its personality? Can you interrupt it? Is it fun to have a "relationship" with? 30. What can it do automatically that computers couldn't do last year. Like does it recognize that I'm aiming at a beautiful sunset and does it automatically capture a photo? How will you judge your personal AIs?

@JosephGruber @FestaInDC I’d place DS9 at the bottom of the list for ST series. Even Picard was a better series.

Come and join in the fun at UCF Engineering for a hackathon with cash prizes +$20k! #ForTheHacker: Reimagine the Future of Law will be a challenge for tomorrow’s talent - how can a law firm provide more value for clients via technology and AI? tinyurl.com/ForTheHacker

Morgan & Morgan @forthepeople

3 years ago

🚨 Programmers, lawyers, data analysts, and tech enthusiasts! Join us at @UCF for a hackathon and cash prizes totaling over $20k! Let's reimagine the future of law. Sign up: forms.gle/d7uLreY6zvJQDL…

1 3 13 4K 1

@elonmusk Spam. Lots of spam.

@SpaceX 🤝 #forthepeople @SpaceX

ChatGPT is coming soon to the Azure OpenAI Service, which is now generally available, as we help customers apply the world’s most advanced AI models to their own business imperatives. azure.microsoft.com/en-us/blog/gen…

#ChatGPT is a powerful tool, but it's just the beginning. Imagine the limitless possibilities of AI-driven technology in our hands. Let's use it to create meaningful connections, solve complex problems, &make the world a better place. #AI #innovation #motivation (Written by AI).

AI is disruptive in enterprise software because for decades innovation has been making people more efficient in how they *already* work. AI can fundamentally change the nature of the underlying work itself, which creates incredible new opportunities.

Feeling incredibly proud repping Morgan & Morgan. We've won millions this month for our clients! #ForThePeople #MorganGotMore #MorganMath #LAW

Morgan & Morgan @forthepeople

4 years ago

Proud to share our latest verdicts! Fighting #ForThePeople is why we do what we do 💪 We have recovered Millions for our clients this past month!! #MorganGotMore #MorganMath #LAW

1 15 6 0 0

@HackForumsNet @elonmusk Russian bots

It's official! Our 2023 @MattMorganESQ Walk For Wishes registration site is LIVE! Thank you to @4RSmokehouse for donating lunch at the Committee Kickoff! Registration is FREE until the end of October and you can register as a team or individual here: orlandowalkforwishes.org.💙⭐

keyboard_arrow_left Previous keyboard_arrow_right Next

@cioontherun

Morgan & Morgan @forthepeople

4 years ago

We are proud to be an official partner & the first legal practice sponsor of the Boston #RedSox!

3 8 22 0 0

I hope that even my worst critics remain on Twitter, because that is what free speech means

Twitter accepts buyout, giving Elon Musk total control of the company theverge.com/2022/4/25/2302…

Thinking about a new car and wish I could see how much I’ve used FSD vs EAP features on my Model 3, @elonmusk can you add reporting?

@elonmusk @wholemars I drove 1/2 the length of Florida yesterday and only needed about 8 minutes to charge to reach Miami. I had to extend my charge time inside @Wawa so I could get something to eat.