Sn0wCoder

Not going to lie I was exited to see this work lol. Thanks for the tip of the day. I love learning and not sure what I was expecting to happen when I opened up excel and hit alt + f11 but it was awesome to see it open. Already created my first user form and now the rest of the weekend its going to be difficult not to go down this rabbit hole. VB once paid my bills....

Looks like its still in the discussion phase, did not see a timeline but also did not read super close. Did learn about the tanstack-router which I had never heard of, so that was worth looking this up right there ;)

https://github.com/solidjs/solid/discussions/2425

Going to tell the Preface story tomorrow at work, I guess if your reading this you now know my HN handle ;) Not sure I will do the 'game' but that into was worth clicking the link.

Watched the video and super excited to see where tech like this ends up in a few years. Was on the quest to find a split keyboard last week and ran across this. Ended up with a UHK, but man I thought hard about buying one (pair) of these. I was just worried about dust, dirt and overall longevity. Seems like something that would need to be worked on every month or two. The kit is reasonable and might be a fun (next) project.

https://svalboard.com/

Would be interested to hear about anyone's experience that owns one of these. How do they hold up? How long did it take to get used to using? Would you recommend to a friend?

Hi Michael, thank you for all the hard work that goes into the Phind models and congratulations on the new UI. Been a paying customer since first finding Phind here on HN in September.

Question: are there any plans to allow access via API to integrate the Phind models with the Continue plug-in (would also love to integrate into my personal RAGs)? Mostly using IntelliJ and integration would be awesome. Do have the VS Code plugin setup and use that when needed. Also running the smaller Phind models locally to use with Continue, but that only gets me so far without needing to open the real UI. If the API opened both the 405B for chat and the 70B for auto complete would be a big step in gaining more paying customers (IMO). No need to open the other models as those can be done with other API keys (if one wanted).

If there are no plans to open the models via API are there plans to support other IDEs (IntelliJ) with the chat feature?

Please let us know!

The Paper.js library is neat if you like this site (found it looking through the source). http://paperjs.org/

You have to pay if you want to use the full API but you can chat with it for free https://www.deepseek.com/

If anyone is trying to run these models (DeepSeek-R1-xxx) on LM Studio you need to update to 0.3.7 Was trying all day to find the error in the Jinja template and was able to make them work by switching to manual then in my email see they added support in the latest version. It was a good learning experience have never really needed to fiddle with any of those settings as most the time they just work. If you did fiddle with the prompt hitting the trash can will restore the original and once you upgrade the Jinja parsing errors go away. Cheers!

I have not used Mixxx nor do I have it installed, but might download to check it out.

But..... If its always in a SQLite DB you should be able to use DB Browser for SQLite to inspect the DB Schema and then write a bash/python (whatever) script to pull the info out.

RESULTS=$(sqlite-utils "data.db" " SELECT song, artist, duration, FROM my_table WHERE song = 'CURRENT';")

RESULTS=$(sqlite3 data.db <<EOF SELECT song, artist, duration, FROM my_table WHERE song = 'CURRENT'; EOF )

I have no idea what the schema looks like but those are just some examples of how straightforward it might be. Run it on CRON (whatever) update as needed. https://cronitor.io/guides/python-cron-jobs

Once you have the Schema an LLM could most likely do the rest if you are not a programmer, but still need someone to get it added to the site.

Would guess GGUF so you can run on llama.cpp, LM Studio, etc..., but OP can hopefully clarity further for you.

This is a fair point to bring up.

Most sites will have a confirmation once you click the link that includes the browser version and IP address. I have seen that info only in the email itself too with no confirmation afterwords, but not for some time. Have never seen one that is just a link with nothing else that once clicked allows the other device in but supposes could be implemented that way.

The article itself is about not making them the only option (which is fair), and the OP says if they do it should login the device which originally made the request (which I agree). If the implementation is just an email with only a link, no other information with no confirmation (yes, it's fine to let this device in), then I would have to agree with you it's very risky and could allow anyone to login as you (hopefully no sites are doing this, but...)

Agreed! If they all worked like this would be a happy camper. Nothing worse than being in one browser, opening the email, then it opens and authenticates you on the default browser or even better on a different device and needing to forward the link to the other device so you can open it there (yes odd scenario but try not to access certain emails from certain devices).

Sites that send an OTP (crazy-pink-horse-3837) that you can copy, and paste is a good middle ground if implementing the link that just Auths the original request is too difficult.

If you have a newer-ish NVIDIA video card you can try ChatRTX. RTX30 or 40 and 8GB VRAM. Its not great but does allow you to point at a folder and is better than nothing. Does docs (txt, pdf, docx) and images (jpg, png, gif) but not at the same time. https://www.nvidia.com/en-us/ai-on-rtx/chatrtx/

Happy New Year may the next be the best one yet....

My partner and I had no intention of having children and we agreed on that. Eventually we moved in together and got married. We both had good jobs, traveled, ski / snowboard, hiked, camped, went to see everything we wanted to see. Long story short we stopped NOT trying to have kids (stopped all birth control) and never happened. Started the whole process of finding out who was the issue, and that was starting to add up quick in the expense department. We were about to start an even more expensive process when an opportunity presented itself that led to us adopting our first child. In less than a year we found out we had conceived one of our own. Sometimes it takes one to jump start the process, I guess. Even longer story short our relationship was starting to drift and then with our children we grew close again. Now we have two more best friends in our family. Coach their teams, host their parties, uber them around, and they are like having two more best friends that live with us. Going to be sad when they leave us but also super excited to be just the two of us again.

This goes to what you are saying about purpose in life. Without them I am not sure we would still be happy or even together. Children gave us a new reason to live and something to work on together. Life is no longer how the two of us can be happy but how we can make others happy, thus giving us pride to be happy ourselves. Strange to think about where we would be without them, but glad I only must speculate not actually know.

I grew up more remote with less opportunity for friends. My partner grew up in a densely populated area with lots of friends (my partner is way more socially adjusted than I am). The ones I made are the ones I still have while my partners are mostly gone and live elsewhere. We decided to raise our family in a large neighborhood with lots of friend opportunities as I feel that is the best way to let them experience life. Personally, I would love to be on 20 acres with my arsenal and survival gear, but that would not be fair to anyone in my family. We got the house with woods in the backyard (and a highway behind that) and both sides for a half acre in each direction, with close neighbors to the front. The rest of the neighborhood is house to house like any other. I need to be able to walk out my back door and take a leak if needed and be alone. While walking out the front provides the typical suburban experience. All comes back to compromise.

Whatever you decide, best of luck! Living the Mad Max lifestyle is hard to beat....

I use IntelliJ as my main coding tool but also use VSCode and Sublime text. If you have access to local LLMs or have an API key for some the Continue Plugin (basically Cursor but can use in IntelliJ) is the Best of the Best for IntelliJ (IMO). I have a box running some local models including Phind and StarCoder (plus some small embeddings) and have been super happy with the end product. The next up is Google Gemini Code Assist has been the best of the IntelliJ (non-configured) AI tools I have tried. There are better ones out there but IMO not for IntelliJ. It's still free for a few more weeks and I have been using it since the free release, fun to use. Can pre-prompt, say you are an expert XXX, please be funny, fill in the rest of your regular prompts. The Co-Pilot I use for work is very limited and will only answer coding questions. I tried to tell it that it was my coding buddy, and its name was Phil and told me it cannot have a personality or be funny. I believe the paid personal Co-Pilot allows you to choose which LLM it uses (I cannot confirm). The Phind VSCode plugin works really well. Also, the Phind coding models are on par with some of the other big ones and free if you have a subscription (or run locally). Sublime is around to open those GIG+ files as VSCode chocks and not worth the RAM of opening another IntelliJ.

Each task / programming language / query requires trying different LLM models and novel ways of prompting. If it's not work-related (or work pays for the one you use) sending as much of the code as relevant also helps the answers be more useful.

Most of the people I meet that say LLMs are not useful have only tried one (flavor / plugin), do not know how to pre-prompt or prompt, and do not give the tools a chance. Try one or two things, say yep, it's not good and give up.

Still hard for me to admit that Prompt Engineering is a profession, but it's the same as Google Fu. Once you learn it you can become an LLM Ninja!

I do not believe LLMs are coming for my job (just yet) but do believe they are going to be able to replace some people, are useful and those that do not use them will be at a disadvantage.

Hey @scyzoryk_xyz I laughed out loud and could understand the tongue in cheek sediment before the edits. When I was younger dreamed about becoming rich and buying a whole subdivision and moving my family and friends into the compound. Obviously, that was before I actually had a wife and children. Compromise is the only way relationships work with family and friends.

Good thing is you have more than enough HN Karma to burn so don't worry about the down votes, most people will get ‘it’ and gave you an upvote to offset the ones who did not.

Thanks for the Christmas fun! Was not going to click as thought it was like an actual snowflake maker (like you know each one is unique generator), glad I did. Could actually make a real one if I was feeling like making a mess ;)

Sorry for your loss. Hope they lived a great life, and you can make it to the funeral.

Without doxxing myself or getting too deep into what I actually do there are some key points to think about.

For very simple to mid complex tasks, I do think LLMs will be very useful to programmers to build more efficiently. Maybe even the average joe is building scheduling apps and games that are ok to play.

For business apps I just do not see how an LLM could do the job. In theory you could have it build out all the little, tiny parts and put them together into a Rube Goldberg machine (yes that is what I do now lol) but when it breaks not sure the LLM would have a prompt big enough to feed the entire system into to fix itself.

Think of this theoretical app.

This app takes data from some upstream processes. This data is not just from one process but several that are all very similar but never the same. Even when from the same process new data can be added to the input, sometimes without even consulting the app. Now this app needs to take this data and turn it into something useful but when it doesn't it needs to somehow log this info and get it back to the upstream app to fix (or maybe it's a feature request for this app). The users want to be able to see the data in many ways and just when you get it right, they want to see it another way. They need to see that new data that no one even knew was coming. To fill in the data that comes in this app needs to talk to 20 different APIs. Those APIs are constantly changing. The data this app takes in needs to send it to 20 different APIs through a shared model, but that model also takes into account unknown data. The app also sends the data from the upstream process to native apps running on the users' local machines. When any of this fails the logs are spread out over on-prem and hosting locations. Sometimes in a DB and sometimes in log files. Now this app needs to run on-prem but also on Azure and also on GCP. It also uses a combination of Azure AD and some other auth mechanisms. During the build the deploy process needs to get the DB, Client Secretes and roles out of the vault somewhere. Someone needs to setup all these roles, secrets, ids. Someone needs to setup all the deploy scripts. Now with every deploy / build there are automated tools to scan for code quality and security vulnerabilities. If they are found someone needs to stop what they are doing and fix the issue. Someone needs to test all this every time a change is made and then go to some meetings to get approval. Someone needs to send out any down time notices and make sure the times all work. Also, the 20 apps this one talks to are the same as this app.

I could keep going but I think you get the point. Coding is only ¼ of everything involved in software. Sure, coding will get faster, but the jobs associated with it are not going away. Requirements are crap if you even get requirements.

The way to future proof your job is get in somewhere where the process is f-ed up but in a job security type f-ed up. If your job is writing HTML and CSS by hand and that is all you do, then you may need to start looking. If your job has any kind of process around it, I would not worry for another 10 – 20 years, then we will need to reassess, but again it is just not something to worry about in my opinion.

I also know some people are building simple apps and making a living and the best thing to do is embrace the suck and milk it while you can. The good thing is if LLMs really get good like good you will be building super complex apps in no time. The world is going to have a lot more APPs that if for sure, will anyone use them is the question.

Think you are being downvoted because technically an LLM stands for Large Language Models, while image generation is something like LDM latent diffusion models or most people just say Stable Diffusion. If you would have said the more generic ChatGPT then the answer would have been maybe since its supposed to be Multi-Modal. If you are asking if Stable Diffusion could generate pictures based of the article description the answer is most likely yes given the prompt correctly parses the context correctly (maybe using an LLM for getting the prompt before feeding into the next tool).

I hear almost everyone call everything ChatGPT no matter what they are using (MS, Google Gemini, Ollama, etc…) ChatGPT made me a PP, ChatGPT made this image, so maybe if you want a general term ChatGPT would be more correct. I typically say ChatGPT for LLMs and Stable Diffusion when talking to normal people since most are familiar with the big two. Predicting the next pixel to draw is sorta like predicting the next word (token) I guess

Most people know what you mean, but still good to use the correct term. Disclaimer so I do not get downvoted with you: Please do your own research as I just wrote how I understand these concepts and am not an expert.

Same issue need to turn off VPN / AdBlock....

The UI is very nice. Congratulations on the release, with all the articles about releasing the last few days must feel good.

Does not like my VPN / AdBlocker, getting a few CORS errors connecting to the Amazon S3. If you really want people to use the site would recommend proxying the request through the server. If you are already running a node.js server straightforward to do, if not still a huge leap but would also want to configure Caddy (NGINX) [or run on the same server and block the port] to run the proxy locally and forward only your requests so it does not get abused.

A lot of the audience here is running AdBlock and / or VPN so others are most likely to hit the same issue.

Watched a documentary on one of the first people producing the lab grown diamonds for rings (jewelry) and he ended up needing to go to China where they were already doing the process for drilling and sandpaper type applications. His process was to add flaws of the right type as to make it almost impossible to tell the difference between lab and non-lab diamonds. From what I can tell from it, about half of diamonds being sold as non-lab are actually lab. When we purchased lab was not big yet so guessing its non-lab (not that it matters other than might have over paid). If we (I) were to do it today would just do lab and be done with it as to ensure did not overpay and end up getting lab anyway. From what I can tell no one can tell the difference other than tiny (id markings/ flaws) that can be faked.

I know there are sellers that claim to know the provenance of the diamonds (can trace them back to the mine), but that would be the best place to introduce them into the chain. For every truck that goes out half could come from another truck and half from the ground.

Ok spent way too much time fiddling with the options and testing out different sort algorithms. It’s fun to start with one sort and then halfway through switch to one that is very different and see if it helps or hurts the performance. Almost makes me want to break out the old Algorithms and Data structures book to remind myself of which one to use when you know the data could already be half sorted vs when you know the data is going to be completely random.

Bookmarked but guessing I will never use it again but who knows if I ever need to do another leet code test this would be a good resource for sorting. This could be useful for those of you that are still in college to have one of those aha moments when it comes to the sorting algorithms you are learning about. If you are going to look at the code, go to src -> lib -> sort-algorithms to see how they are implemented. Note the asterisk after the function* keyword which signifies a generator function.

LOL, I passed by the article a few times reading the same thing. When I finally clicked in thinking it was something to do with the windows package manager. Something about the capital E. As soon as it was spelled with the little e, it clicked.

Great point thanks for calling USING out. Since you end up putting table_name_id for FKs it totally makes sense to just use that in the main table. Seems I am just so accustomed to having id as the default PK over the years it become habit (my DB professor was an old time IBM-er who preached all tables will have an ID). With auto complete in just about every tool these days and ORM limitations improving will need to update my thinking on this reality. 95% of the time living in the MSSQL world so USING is not something that can even be used (I don’t think).

If the table names are all short one word like that, they are already basically aliases. What do you do when you end up on some legacy project (you did not make the schema) where_the_table_names_look_like_this? Seems some sort of alias might be more appropriate. Also, when you are writing longer / sub queries or using Common Table Expressions it’s impossible to not use an alias.

Not bad advice. The one about “where possible avoid simply using id as the primary identifier for the table” stood out to me. In the past with multiple ORMs (ya, ya, we all hate them) the default was to map to a column named id. Also when doing joins its cleaner to use the table_name.id or alias.id then table_name.table_name_id or alias.table_name_id or whatever else besides id is used. The best is when multiple people have worked on the project over the years and the columns are a combo of camel, snake, camel_snake, all UPPER / lower. Must look at the table definitions or ERD every time you want to write some non-trivial query. So having a consistent style guide is better than having any one specific style guide. This would be a good starting point and adjust with your team as needed.