sixdimensional

Not OP and just guessing, but probably SXM2 GPU modules for the V100. Those can be acquired fairly inexpensively, but there is work to do to get them working together and the V100 has some limitations on the types of models you can run.

Astral, bun, Stainless, Cursor... and more..

Seems like developer tools/tooling are a hot commodity to the current big AI companies?

First one to vibe code a language for LLMs, by LLMs, wins a cookie?

I second the idea to use a zip. In fact it's what a lot of vendors do because it is so ubiquitous, even Microsoft for example - the "open Microsoft Office XML document format" is just a zip file containing a bunch of folders and XML files.

A few thoughts from instant gut reaction:

- you're on HN, so you have an opportunity to tell us the name of the thing - marketing

- perhaps consider your marketing budget to be the area you need to invest in now, and indirectly how that budget can actually generate a little revenue and exercise the engine - e.g. do you have packages you can ship through your service to announce it to others? Use your marketing budget to do it and collect your marketplace fees - marketing again. Nothing says I believe in my product like using it for real (IMHO).

- a bold and risky move (?) - most would say fake it till you make it - but I for one tire of this tactic. How about approaching this with honesty and reward the first adopters? "We're brand new but you can be the first to help us prove out this model?".

- your marketplace is a network. Search for an opening that has viral properties. Try to tap into something that has a network effect, go to where your customer is and see how you can target/advertise strategically and respectfully - become a trusted partner to one or more communities (e.g. thinking out loud, eBay sellers maybe?). This could include finding the right partner(s) who have a problem and are willing to give you a shot in an existing network.

On the last point - as an example of the viral thing - I worked on a real estate tool a while back. We found a viral hook - there were for sure properties that needed to be processed and worked through the tool - we email invited the parties involved in the transaction to invite them to work on the property in the secure tool, and we gave them the ability to invite others working the same transaction to the tool, and we focused everything on polishing that workflow and experience.

This way, as soon as one person used the tool they could invite others to use it legitimately to work in the tool and that was the viral aspect.

This points to, replace real estate tool and house with "package" and invite... and can you achieve something viral that spreads itself... like, when someone ships with your tool, it emails the recipient with the link to your site for status tracking and a call to action to make them want to ship using your platform.

This to me is a lot of marketing and product strategy around incentivizing the network effect.

Disclaimer: I never made millions of dollars off a marketplace. But I did help stand up the real estate mechanism I mentioned and that business reliably brought in 5-10 grand a month with no marketing effort and just that one mechanism, and it also helped us find a few key network partners. That's what drives my feedback.

I have a bunch of TI-99 hardware in storage, have been thinking to donate it to a computer museum potentially. I had one in my hand when I was 5 thanks to my grandpa (it made me what I am today!).

Anyone up for a rousing game of Pole Position?

I still don't understand what happened to using Apache Avro [1] for row-oriented fast write use cases.

I think by now a lot of people know you can write to Avro and compact to Parquet, and that is a key area of development. I'm not sure of a great solution yet.

Apache Iceberg tables can sit on top of Avro files as one of the storage engines/formats, in addition to Parquet or even the old ORC format.

Apache Hudi[2] was looking into HTAP capabilities - writing in row store, and compacting or merge on read into column store in the background so you can get the best of both worlds. I don't know where they've ended up.

[1] https://avro.apache.org/

[2] https://hudi.apache.org/

I concur.

Also, it was paid for by US taxpayer dollars - the entire content should have been released somewhere for free, maybe even someone would have started up a new project to maintain it, for example, something under Wikimedia or some other nonprofit.

This wholesale elimination of valuable information and data owned by the public is so incredibly sad and damaging to our future.

Maybe we need a FOIA request to get the entire contents released to the public.

I would enjoy discussion with whoever voted this down - why did you?

What is your opinion and did you vote this down because you think it's silly, dangerous or you don't agree?

Half serious - but is that really so different than many apps written by humans?

I've worked on "legacy systems" written 30 to 45 years ago (or more) and still running today (things like green-screen apps written in Pick/Basic, Cobol, etc.). Some of them were written once and subsystems replaced, but some of it is original code.

In systems written in the last.. say, 10 to 20 years, I've seen them undergo drastic rates of change, sometimes full rewrites every few years. This seemed to go hand-in-hand with the rise of agile development (not condemning nor approving of it) - where rapid rates of change were expected.. and often the tech the system was written in was changing rapidly also.

In hardware engineering, I personally also saw a huge move to more frequent design and implementation refreshes to prevent obsolescence issues (some might say this is "planned obsolescence" but it also is done for valid reasons as well).

I think not reading the code anymore TODAY may be a bit premature, but I don't think it's impossible to consider that someday in the nearer than further future, we might be at a point where generative systems have more predictability and maybe even get certified for safety/etc. of the generated code.. leading to truly not reading the code.

I'm not sure it's a good future, or that it's tomorrow, but it might not be beyond the next 20 year timeframe either, it might be sooner.

I like Fastmail with my own domain for personal email, but the reality is nothing is a complete replacement for a Google account, given how tied in it is with auth and the whole Google ecosystem. I still have to use Google for work.

Proton is another one people often suggest. Hey.com sometimes too. No experience with those myself.

There are other options (such as the big guys, iCloud mail or Outlook.com), but aside from self-hosting (which I don't want to spend time maintaining just for my personal mail), I personally haven't seen much outside of those ones that are recommended often.

I do.

Not selling anything, but I am trying to figure out what to do to help support solo and micro entrepreneurs, very small businesses (2-3 people) and very small nonprofits.

I feel like there are a lot more people in this position now (me included), but I don't want to do things for the sake of doing them... I want to find out what solo folks really benefit from and help make sure you get more support.

Thank you for your response!

I agree that spending time on inference or compute every time for the same LLM task is wasteful and the results would be less desirable.

But I don't think the thought experiment should end with that. We can continue to engineer and problem solve the shortcomings of the approach, IMHO.

You provided a good example of an optimization - tool creation.

Trying to keep my mind maximially open - one could think of a "design time" performance at runtime - where the user interacting with the system is describing what they want the first time, and the system is assembling the tool (much like we do now with AI assisted coding, but perhaps without even seeing the code).

Once that piece of the system is working it is persisted so no more inference is required, as essentially code - a tool, that saves time. I am thinking of this as essentially memoizing a function body- i.e. generating and persisting the code.

There could even be some process overseeing the generated code/tool to make sure the quality meets some standard and providing automated iteration, testing, etc if needed.

A big problem is if the LLM never converges to the "right" solution on it's on (e.g. the right tool to generate the HTML from the SQL query, without any hallucination). But, I am willing to momentarily punt on that problem as being more to do with the determinism problem and the quality of the result. The issue isn't per se the non-deterministic results of an LLM anyway, it's the quality of the result fit for purpose for the use case.

I think it's difficult but possible to go further with the thought experiment. A system that "builds itself" at runtime, but persists what it builds, based on user interaction and prompting when the result is satisfactory...

I remember one of the first computer science things I learned- the program that could print out it's own source code. Even then we were believing that systems could build themselves and grow themselves.

So my ask would be to look beyond the initial challenge of the first time costs of generating the tool/code and solve that by persisting a suitable result.

What challenge or problem comes next in this idea?

I totally agree. The reason I asked before offering any solution ideas was I was curious what you might think.

My brain went to the concept of memoization that we use to speed up function calls for common cases.

If you had a proxy that sat in front of the LLM and cached deterministic responses for inputs, with some way to maybe even give feedback when a response is satisfactory.. this could be a building block for a runtime design mode or something like that.

Have you tried the thought experiment though?

I agree this way seems "wrong", but try putting on your engineering hat and ask what would you change to make it right?

I think that is a very interesting thread to tug on.

I think your last comment hints at the possibility- runtime generated and persisted code... e.g. the first time you call a function that doesn't exist, it persists if it fulfills the requirement... and so the next time you just call the materialized function.

Of course the generated code might not work in all cases or scenarios, or may have to be generated multiple times, and yes it would be slower the first time.. but subsequent invocation would just be the code that was generated.

I'm trying to imagine what this looks like practically.. it's a system that writes itself as you use it? I feel like there is a thread to tug on there actually.

This brings a whole new meaning to "memoizing", if we just let the LLM be a function.

In fact, this thought has been percolating in the back of my mind but I don't know how to process it:

If LLMs were perfectly deterministic - e.g. for the same input we get the same output - and we actually started memoizing results for input sets by materializing them - what would that start to resemble?

I feel as though such a thing might start to resemble the source information the model was trained on. The fact that the model compresses all the possibilities into a limited space is exactly what makes it more valuable - instead of having to store every input, function body and outputs by memoizing that an LLM could generate, it just stores the model.

But this blows my mind somehow because if we DID store all the "working" pathways, what would that knowledgebase effectively represent and how would intellectual property work anymore in that case?

Thinking about functional programming, to me the potential to think of the LLM as the "anything" function, where a deterministic seed and input always produces the same output, with a knowledgebase of pregenererated outputs to use to speed up the retrieval of acceptable results for a given seed and set of inputs.... I can't put my finger on it.. is it a basically just a search engine then?

Let me try another way...

If I have a ask an LLM to generate a function for "what color is the fruit @fruit?", where fruit is the variable, and I memoize that @fruit = banana + seed 3 is "yellow", then the set of the prompt, input "@fruit", seed = 3, output = "yellow"... then this is now a fact that I could just memoize.

Would that be faster to retrieve the memoized result than calculating the result via the LLM?

And, what do we do with the thought that that set of information is "always true" with regards to intellectual property?

I honestly don't know yet.

Maybe something here might help with masonry stuff: https://css-tricks.com/piecing-together-approaches-for-a-css...

I stumbled across it looking for CSS flex masonry examples.

There is a good solution to concerns over this...

Use UUIDv7 for your primary key but only for large scale databases and only for internal keys. Don't expose these keys to anything outside the database or application.

Use UUIDv4 columns with unique indexes over them as "external IDs" which are what is exposed via APIs to other systems.

Basically, create two IDs for one record - one random but not the primary key, and one sequential, that is the primary key.

I have done this in real systems.. and it works.

If you're writing out data for analytical read only use - go with DuckDb all the way, over SQLite.

I'm not sure quantum computing is the solution, but it strikes me that a completely new compute paradigm like quantum computing is probably what is necessary - which is orders of magnitude more efficient and powerful than today's binary compute.

I feel like the fundamental concept of symbolic logic[1] as a means of reasoning fits within the capabilities of LLMs.

Whether it's a mirage or not, the ability to produce a symbolically logical result that has valuable meaning seems real enough to me.

Especially since most meaning is assigned by humans onto the world... so too can we choose to assign meaning (or not) to the output of a chain of symbolic logic processing?

Edit: maybe it is not so much that an LLM calculates/evaluates the result of symbolic logic as it is that it "follows" the pattern of logic encoded into the model.

[1] https://en.wikipedia.org/wiki/Logic

I'm not sure I totally agree.

Personal profit maximization only works to a point - for example, if you get too old, sick or the system rejects you early and curtails or limits your ability to make money.

I don't disagree that money gives you options, but, far too many people wait until they have enough money to give back.

If you give back while you are working (e.g. balancing working for profit vs working for nonprofit, altruistic reasons, etc.) - that's awesome. The challenge there is maximizing the good you can do if you're giving too much time and energy to your profit maximization.

At some point, someone has do physically do the needed good work.

For myself, the calculus has shifted. I personally decided I cannot wait until I have enough money, or I am maximizing my profit, to go out and help people.

I also cannot wait until I am physically or mentally unable to help beyond financial contributions. Also, I cannot afford to work in the current system that drains everything from you and leaves you no energy or time left, only money (if that).

Regarding the inherent maximum scaling limits of one person- I would challenge your thinking.

Power laws of networks may demonstrate that helping a small number of the right people might be enough to unleash the butterfly effect or play into ongoing changes.

Also, the physical limits of humanity on one person apply to a billionaire as much as a person with little money. I'm not saying a billionaire, millionaire, or person with significant finances isn't more mobile/capable, but it's not a given.

I am for reasonable profit and balance. There is nothing inherently wrong with maximizing profit if someone chooses.

But if we all spend our time on maximizing profit, there still, for the time being and probably well into the future, still needs to be boots on the ground doing work that is not for profit.

Sorry to here about your bad experience. I do understand it, as when the hype cycle causes perfectly good things to be thrown away or messed up for no practical reason, it really bugs me too.

In this case, it's so easy for some to say everything else is bad if it is not AI, or you have to include AI "because it's the future".

I just wanted to clarify that I am very pragmatic in my approach to new tech and AI.

I used to work on enterprise ERP, MRP, sales and support systems (Oracle, Salesforce, ServiceNow, and more)... not always by choice.

At one point a project came up to rebuild our core customer experience portal across our business. I got to build with the rule based chatbots to implement an assistant in that web app.

Not easy, but it worked well enough, especially for simple scenarios. For example, imagine a field support engineer being able to pull up the exact page in a big technical document for some obscure part by just asking for what they need, in seconds.

Or, being able to reorder some quantity of part by just asking for it, or checking the quantity of same.

These are time savers, especially if you have a voice to assistant interface too. So you can just type something, or ask for what you need without typing.

Tablets were common for manufacturing, inventory or service staff. Being able to pull up our interface and just say what they needed and get prompted for simple things - there were TONs of pragmatic small wins there.

Current LLM just make that easier than the old rigid rule based way we had to code those assistants (at the price of going wrong sometimes).

I care about efficiency, pleasant user experience, and pragmatics.. not AI for the sake of AI.

All that to say, if a drop down box works better for some use case, instead of AI - or even a rule based chatbot - hell yeah! Save on them tokens.

I somewhat agree with you, especially that one could identify a common abstraction that later an LLM could piggyback on top of.

Genuine question though - have you implemented an AI assistant/chat interface recently using LLMs on top of a UI?

I agree it can be a rabbit hole, but I just got through doing it on an app and there were definitely some things it really made way simpler and some complex scenarios that I'm not sure could have been done any more simply.

Cool! Yeah, that's the kind of UI/UX I meant.

I agree with the right abstraction and it's tough to find the balance- in our data pipeline app, what we did is make key core functionality of the app exposed so the assistant can use it, and implemented a handful of basic agents out of the box, including one default one that could shell out work to others. We also made it easy as an extension point for users to add a new agent that used the core functionality/tools, just by defining the agent in a markdown file.

We found starting small for critical use cases that saved the most time, but thinking about building blocks, was useful.

Because the responses of the AI assistant come back and are processed on the UI, we found we could give the LLM our UI docs as well as knowledge about UI element IDs, etc so it could respond with input commands that would drive the UI.

This way, we could do something like, provide the LLM with the input/prompt including the context of like - what page/view is the user on, what is their intent, what tools are available in general, what sub agents are available for specialized tasks, etc.

Please don't let my suggestions sway you away from core progress in your app (take with a grain of salt). But it's great you're already experimenting- keep your eyes open if you see a great use case where it accelerates workflow.

Another HNer mentioned people not reading docs- that's a low hanging fruit use case we had too - "how do I use this view?", "what does this field mean?", or retrieving information from other parts of the app without having to navigate away, etc. It can save having to find answers in a doc or navigate elsewhere.

Edit: perhaps a useful exercise - imagine a workflow of "talking to the app to achieve a task" as a way to explore.

"Hey ERP, open the part entry screen for part 12345"

"Hey ERP, can you update the description for part 12345 to correct the spelling error?"

"Hey ERP, how many of widget XYZ are in stock? If there are enough in stock, can you transfer quantity 10 from warehouse A to B?"

"Hey ERP, how do I cancel a sales order?"

"Hey ERP, how does this screen work?"

I think if you break these down, you'll find common abstractions that map to features, API endpoints, user interface sequences and interactions, triggering workflows, looking things up in docs, etc.

I agree with you.

I used to work in green screen text based UIs from the 80s (TUI).

Power users didn't need search or anything else - they memorized the keystrokes to navigate the text UI and could just type key combinations to blaze through the UI faster than it could render on the screen.

I've never really seen anyone able to replicate that UI in a browser based or GUI desktop app to be honest.

Power users are a different use case, although the end goal is to remove the need for clicks.

I don't believe chatbots/AI assistants are a panacea, definitely encourage the architects of this new ERP platform to weigh the pros and cons.

That said, two jobs back I worked for a major manufacturing company that used the old GUI destop based Oracle EBS ERP. To automate repetitive workflows they were trying to implement UIPath (RPA automation - it drives the UI for the user) on top of the GUI.

This is what lead me to believe that if the ERP application's functionality is discoverable by an AI assistant, it can be used to automate or navigate on behalf of the user, or as part of complex workflows.

That can be done later after the basics are addressed - my only advice would be to just consider it sooner, even if you don't build it first.

It's a little easier to think of how one might simplify the workflows and design for automation into the core of the product, via the UI, APIs, etc earlier than later.

But in general - focus on the user's needs first and different roles/personas - just don't completely ignore new types of automation workflow opportunities (i.e. AI assistants/chatbots).

My opinion only.

No, I mean, like a copilot style AI assistant, the user can chat with to ask what they want to do, and either the assistant can operate and navigate the UI to the right place, or perhaps even shortcut the steps for the user by asking for questions to satisfy inputs for the thing they are trying to accomplish.

An example: - user intent is to update an attribute for a component part number A21445

- user can click a chat bubble icon in lower right and chat to the assistant

- user describes their intent - "help me update the description for part number A21445

- system replies by informing the user it will open the right screen, opening a part/component editing UI, with the right part loaded, with the cursor positioned in the description field, and the assistant stays open for further assistance; or;

- system replies that it found the part and can update the description, shows the current description for the user, asks "what description do you want?"

- user enters updated description

- system confirms the change is correct

- user confirms the change is correct

- part/component description is updated without even opening the UI

FWIW, it's great that you have cmd-K and also I've seen those kind of search boxes get more smarts like being able to type "part:A21445" to go directly to a specific UI.

I just suggested the above as we learned some interesting user experiences became possible when our AI assistant had the ability to control our UI directly on behalf of a specific user.

An example in the app I worked in (a web based data pipeline tool):

- "Hey assistant, can you help me add some SQL transformation logic to dataflow ABC, to process the customer data?"

- system uses metadata and knowledge of the UI to open the right dataflow, select the right type of UI to open to enable the user to add a SQL query in the right place, maybe even autogenerate the initial SQL query - this all from the main home page of the app, from a side panel chat assistant.

- net result feels like talking to the assistant to operate the app, almost no clicks required.

I hope that makes sense.

Give users an AI assistant they can ask to navigate them to the right screen or section of the application?

In a previous job, we built our AI assistant so that it could operate our UI in the front-end and it was very powerful.

It wasn't an ad. I used to work for the parent company of PWS years ago, and their weather model was really awesome.

I thought the TAMDAR tech reference might be if interest to those following and reading comments here.

I'm not associated with them anymore and I'm not sure how much of their model still exists after the assets were sold to flyht, but I hope it does because we need good weather models!

Fair enough being vigilant, but not an ad.