I've noticed there seems to be a default style that is easier to detect. I've noticed it harder to detect when asking an LLM to use a different style (more conversational, avoid sounding like an AI, don't use emdashes, etc). I wonder if that's what you're picking up too - the instances where people make no effort to change the style of the output.
As a long-time Costco member and very minor shareholder (like 10 shares), lawsuits like this are frustrating. It is in my best interest as both member and shareholder for Costco to relentlessly look for opportunities to reduce costs, including getting credits back from procurement and sourcing. It would be costly to try and determine the tariff impact to every member and then pass it back along. I'd rather see those funds contribute to keeping prices low by offsetting other cost pressures.
I have 3 kids (14, 12, 11). I could have used something like this when they were around 10-12 to create some accountability for simple tasks like "pick up your room" or more commonly the dreaded 20 minutes of reading, especially when they ask for extra screen time.
The demo video has more than I would ask my kids to do, but 1-2 minor things that should be done every day isn't a sheesh to me.
It seems the problem is looking for a single picture to represent the whole. Why not have generative AI always generate multiple images (or a collage) that are forced to be different? Only after that collage has been generated can the user choose to generate a single image.
I have been looking for a solution where I can remote into my company laptop in a pinch from my phone. Many of our cloud resources are moving to be private IPs and VPN only, which can only run on corporate managed devices. I have a few PIs sitting around and it seems like it would be $100-$200 even with a PI to build (pikvm hat is $150 and tinypilot wasn't obvious what the BOM is).
Is there a lower cost or simpler solution? Company hardware is windows and VPN is fortinet and WSL2 is allowed
Looks good. We use CoCalc for similar collaboration benefits. There is a self-hosted option which was important to us. CoCalc has been a game changer as we've all moved remote. Once Deepnote adds the self-hosted / cloud option (I see it coming soon) we'll check it out.
Seems like an advertisement for a book to get access to the content. The book is authored by the same person who wrote the article and the poster is new. I flagged it
I understand that naming conflicts make things harder to search for, but I don't like the notion of a 'we' saying how things can be named. People should be free to name their projects how they see fit and whatever connects with them. A person would likely need to search for 'unicorn emulator' anyways and it comes right up.
There is likely a sizable number of your trial users from HN who signed up merely to better understand it and see if it actually works. It crossed my mind to do it, so presumably others too. The initial spike could be largely driven by technical curiosity instead of real need.
Nice work. A few thoughts came to mind after reading other comments.
I wonder how the tree concept works when responses are long or short. It seems to me that short would encourage more of a "ping pong" style thread which may be desirable in a tree form instead of a long-form question and reply.
It may be interesting and enjoyable to contribute to a site if there was a short length constraint placed on the reply. For sake of example, let's say it was 140 characters. That would force replies to be concise and give more of an anchor for others to jump off on in the conversation.
It could also make it more of a competition -- who can succinctly reply to the question with the most clarity?
Lastly, I think this could go a long way towards mitigating the curse of knowledge when people communicate. I see this and experience this when I participate in technical forums. It's easy to use jargon or advanced topics that the person may or may not understand. The tree approach gives a simple way for a follow-up question around the topic.
You may even consider tying this into the "five whys" which is a common questioning approach
This looks similar in many ways to mithril[1] -- somewhat similar api - mount, view, r("input") vs m("input"). The counter example[2] looks similar[3] I'd be interested in seeing comparisons between the two.
Your templates feature is useful and seems to be missing in most notes apps I've looked at. We are looking for a tool to create our playbooks in and then track the execution of instances of those plays. For example, a "onboarding playbook" can be cloned for each instance of a hire.
Similarly, we and many other companies have their own custom project lifecycle which would be a useful template to be able to create and track for each project.
Yes, I would pay a little more. $5/mo is the max -- any more than that I would probably just do a family plan or something. I would prefer something nominal like $2/per device/per month
I like the idea of this as a second device that has phone capabilities. I wish it were easy/quick to switch between multiple devices on the same cell plan. I don't particularly want to go out running with my google pixel, better to use my 4 year old phone for that. I could see taking the Ektra out on a hike if I could toggle it on as the activated device without going through the hoops of calling verizon.
It seems KDB excels at queries, joins, and aggregates on large datasets. I have limited experience with redis, but if that type of code would need to be written in c/python with redis then KDB may have a performance and productivity lift. I would like to see an example like that.
This greatly helped my understanding of how compilers work.
How different would the implementation be for x64? Would it primarily be a matter of translating opcodes and calling conventions? For example, pop ecx is 59 [1], and on x64 it looks like it would be pop rcx, which is also 59 [2][3].