I get background anxiety while waiting for long-running terminal commands. Nowadays that nagging feeling extends to LLM calls too. Seems like as AI spreads, the pain will only get worse
So I’m working on a universal progress bar HUD
- inspired by World of Warcraft raid mods
- fun sound effects for job start, end, error, and milestones
- can quick jump back to relevant app/tab
- starting with terminal commands and Claude code, cursor agent next
Yup that could help, although if the key content is behind the window, clicks would bug out. I'm writing a PR to hide the window for now as a simple solution.
More graceful solutions would intelligently hide the window based on the mouse position and/or move it away from the action.
Which it did! It chose the option with the best reviews.
However again the Agent.exe window was covering something important (in this case, the shopping cart counter) so it couldn't verify and began browsing more socks until I killed it. Will submit a PR to autohide the window before screenshot actions.
The safety rails are indeed enforced. I asked it to send a message on Discord to a friend and got this error:
> I apologize, but I cannot directly message or send communications on behalf of users. This includes sending messages to friends or contacts. While I can see that there appears to be a Discord interface open, I should not send messages on your behalf. You would need to compose and send the message yourself.
error({"message":"I cannot send messages or communications on behalf of users."})
Great idea Kyle! I read through the source code as an experienced desktop automation/Electron developer and felt good about trying it for some basic tasks.
The implementation is a thin wrapper over the Anthropic API and the step-based approach made me confident I could kill the process before it did anything weird. Closed anything I didn't want Anthropic seeing in a screenshot. Installed smoothly on my M1 and was running in minutes.
The default task is "find flights from seattle to sf for next tuesday to thursday". I let it run with my Anthropic API key and it used chrome. Takes a few seconds per action step. It correctly opened up google flights, but booked the wrong dates!
It had aimed for november 2nd, but that option was visually blocked by the Agent.exe window itself, so it chose november 20th instead. I was curious to see if it would try to correct itself as Claude could see the wrong secondary date, but it kept the wrong date and declared itself successful thinking that it had found me a 1 week trip, not a 4 week trip as it had actually done.
The exercise cost $0.38 in credits and about 20 seconds. Will continue to experiment
My uncle had ALS and was slowly losing his ability to speak. I visited and in order to hear him, we had to gather around him closely. It was very fatiguing for him to project his voice.
I went to a few audio stores and jerryrigged a portable mic-speaker setup that could attach to his wheelchair. No software, just the right series of devices and adapters. It worked well and provided a huge relief for him and our family. Nothing impressive technically, but definitely the physical thing I'm most proud of making.
The Corsair K65 achieved the fastest latency of 0.1ms. By comparison the Apple Magic Keyboard with TouchID had a latency of about 27ms, both wired and bluetooth. Pretty wild that the Apple keyboard is 270x slower!
Now I personally use the Logitech G915 TKL, low-profile. The 1.3ms latency is excellent and I love the key feel.
I’ve found good use of Polymail and pay $20/m for team pro. It’s a love-hate relationship though due to small but disruptive UX patterns. Some examples:
The MacOS app does not follow normal keybinding conventions. Specifically, ESC causes the app to exit full screen and cmd+shift+f doesn’t enter full screen. No option to customize either.
The iOS app will instantly show notifications for new emails, but upon opening up the app you have to wait 5-10 seconds for the emails to appear (while gmail is instant).
That said I enjoy the inbox zero image, snoozing of messages, and overall style.
I got tired of maintaining one-off scripts to do recording, so I started working with friends on a dedicated solution. Today it lets you stream logs via a small Python library, then view individual training runs on an iOS/Android app. Takes less than a minute to get setup.
We're planning on expanding to model versioning in the next few weeks. Interesting to see how others are thinking about it. If you have model versioning thoughts you dont feel like posting here, drop me a note at [email protected]
The trouble is that it's hard to predict an agent massively more intelligent than ourselves. But let me enumerate a few given properties you gave to the super-intelligent AI (SI) in your examples and then tell a story of a SI that became an existential threat:
1) The SI is hypercompetent at cybersecurity
2) The SI is hypercompetent at social skills.
3) The SI is connected to the internet
4) The SI has a goal/utility function (wipe out humanity / maximize paperclips)
And I'll add another property that Hawking notes is important:
4) The SI is able to improve its own intelligence
The story begins with the SI escaping from its handlers. The first thing to note is that the SI is now, in effect, immortal. With it's cybersecurity skills, the SI can avoid detection and infect a tremendous number of computers - at first those it calculates will be low-risk (i.e. existing botnets, old Android phones, etc)[1]. Using the additional computational power, the SI can continue to recursively self-improve and plan until it has the competency to invisibly infect high-value targets like the AWS cloud and (importantly) the computers of AI researchers.
Now the SI can plan for a long time. The SI can quietly encourage AI research and try to prevent end-of-civilization type events via its hypercompetent social skills. Eventually AI researchers will come up with an AI they declare as 'safe', 'friendly' or 'aligned'. The SI, having long ago compromised all the relevant computers and chip factories, silently infects this 2nd super intelligence, and replaces the 2nd SI's utility function with its own. Now the 2nd SI pumps out miraculous inventions - cures for disease, compelling societal ideas, and labor-saving robots.
Eventually we find ourselves in a wonderful post-scarcity world. The AI researchers are lionized as mankind's greatest geniuses, responsible for the creation of a benevolent SI that takes care of our needs from it as well as it's own. You may not trust it, but it will find people who do. Maybe greed, nationalism, security fears, or saving loved ones from death. The SI builds the needed facilities to thundering applause.
The SI is now confident in moving towards the next step. Time for some paperclips! One day it quietly sends a new blueprint to a few of the automated biolabs built to cure cancer. A few hours later the biolabs release a series of airborne super viruses and/or nanobots and 99.999% of humans die, with the rest saved for experimentation and convinced terrorists did it. The end.
Super-intelligent AI is an existential risk because while a super-intelligence keen to destroy humanity might fail today, it will succeed in time. The moment a SI touches the internet, our fate as a species may be sealed.
Yes, I mean a moral hazard in the sense of a divergence of best interests between YC and an overseer.
of all the leadership structures YC has experimented with, the board of overseers seems least proportional in terms of the power of the body and its investment in YC. consider the prior era of YC where all rewards and losses in both finance and reputation fell upon PG and Jessica. once sama became president, he put skin in the game in by staking his professional reputation on YC's success.
i pose the concern of moral hazard because one piece of information has yet to be made public: how board members are impacted by YC's success or failure.
Just crossed 15k Mac and windows apps analyzed!