Ask HN: Who's building an AI-free product? | Hacker News Reader

Ask HN: Who's building an AI-free product?

50 pointsleonagano68 comments

Share your product below

Hey guys! We're engineers/designers from France, and we've built the Ultimate DIY Battery that you can repair and refill!

It works with 90% of the bikes/motor brands on the market, so I assumed that some people here might be interested, if they got a non-functional batteries but they still want to use their e-bike?

We believe that everybody should have control about stuff they own, and we should fight against planned obsolescence!

Here are a few videos about our founder on the battery itself, why we built it, and how to assemble it:

- What is the Gouach Battery: https://www.youtube.com/watch?v=NsuW1NPkvNk

- Presentation of the pack: https://www.youtube.com/watch?v=mLoCihE0eIA

- Presentation of the fireproof and waterproof casing: https://www.youtube.com/watch?v=EDJpt7RDbRM

Here are the juicy bits: https://docs.gouach.com

We'd love some feedback from the e-bike DIY builder community

Oh, and it's launching as a Kickstarter in September and there is an offer for early-backers here https://get.gouach.com/1 for a 25% discount on the battery!

You can follow us on Instagram https://www.instagram.com/gouach.batteries to get the latest news!

Its better to present as 'fire resistant' than 'fire proof' I think

Interesting, but then I saw the price. Ouch.

How are you planning to compete with "asian brands" who see the same thing (less all the "eco" fuss) for a fraction of your price?

We're about $330 with the 25% off, which is in line with most high-quality battery brands (an much cheaper than Bosch or other)

Oh yes, no interest in AI whatsoever. I've been working on Habitat for the last 4 months. It's a self-hosted social platform for local communities. The plan is for it to be federated, but that's a while off yet.

- The idea: https://carlnewton.github.io/posts/location-based-social-net...

- A build update and plan: https://carlnewton.github.io/posts/building-habitat/

- The repository: https://github.com/carlnewton/habitat

- The project board: https://github.com/users/carlnewton/projects/2

i cant seem to run it maybe its skill issue

Hey, I'm sorry, I missed this comment. It's still in early development but if you have experience setting up docker environments, it should be possible to set it up locally. I've put a lot of work into investigating production setup in recent weeks but I haven't yet published any of that, but if you want any help with local setup I'd be happy to problem solve.

We're building WorkMode (https://workmode.net/) - a body-doubling platform for extreme procrastinators. In a nutshell, we connect with our users via video call to ensure they stay focused on their tasks. These video calls usually last the entire day. Think of it as "Focus/Productivity as a Service."

Replacing our productivity partners (body doubles) with AI avatars simply wouldn't work. The method's effectiveness relies on interaction with another human. We asked some of our users about this, and most of them said they wouldn't feel much pressure to work on their tasks if they knew a machine was watching them. In fact, most of them would feel horrible.

I logged in after ages to reply. This sounds useful. I'm having trouble concentrating on my work, remote work has become too much of a toll on my mental health. But that being said, it's so crazy that there's a need for this. How are you guys making money? I see you have a single person tier for 3 USD an hour? Even so, that's crazy.

Thanks! For some, it's the only method that works (myself included).

> How are you guys making money? I see you have a single-person tier for $3 an hour? Even so, that's crazy.

Update: First of all, users are _not_ our product. Our only source of revenue is subscriptions to our service.

Each user has a dedicated Productivity Partner, and they connect with them in a 1-on-1 video call. 90% of the time, Productivity Partners do not interact with users; they just ensure the user is present and (optionally) monitor their screen to make sure they're not scrolling through Reddit. This allows each Productivity Partner to work with multiple users simultaneously.

By the way, screen sharing isn't as scary as it seems. We lower the resolution and bitrate on the user's side to such low levels that we're not able to read any text or even distinguish images. However, it's enough for us to see when the user is on Facebook, YouTube, Reddit, etc.

Edit: more context in quoted message. Edit #2: clarification that you're not our product

This sounds like a nightmare and very questionable in any serious work capacity. You are paying to breach your nda...

Meta: Sigh, miss click - deleted the wrong comment, so here it goes again.

Screen sharing is optional - it's up to the user to decide whether they want to use it.

Not everyone works under NDA. Some of our users are medicine/law students, entrepreneurs, life coaches, history professors etc.

Here is an AI feature for you :D => Users share their screens and you detect webpages, and applications on the client side (process frames/images, detect logo, extract text from it, etc.). Don't send the stream to other parties only you share detected apps/webpages with others.

We can think of several AI applications for WorkMode, but we don't want to implement them. In our case, the less AI, the better.

AI, automatic data gathering, big data processing - no thanks, we'll skip it. What we do works; we don't need to improve it - at least not with AI. We don't want our Productivity Partners to handle hundreds of users simultaneously, as this would result in losing the personal connection with the user.

Sounds similar to https://www.focusmate.com/ - are there any big differences?

The concept is the same, but the execution is different. As a rule of thumb, Focusmate may work well for light, occasional procrastinators, while WorkMode is tailored to the needs of lifelong procrastinators.

I'm going to mention some disadvantages of using Focusmate, but please note that I'm not criticizing their platform. If Focusmate works for you - great, continue using it. However, if you find yourself scrolling through Hacker News instead of working, you might want to consider using WorkMode.

So, here are some differences between WorkMode and Focusmate:

On Focusmate, you're connected to another procrastinator who is there for their own benefit and usually doesn't care about your productivity. On WorkMode, it's different—you work with our employee whose only job is to ensure that you're working on your tasks.

On WorkMode, there's no such thing as a "no-show." I'm a hedonistic procrastinator, and even one missed session can result in skipping work for multiple days. This can be catastrophic if you charge by the hour (e.g., many freelancers and consultants). When I started body doubling, my income rose by 2-3x, and my disposable income increased even more.

We call you every day 30 minutes before your scheduled session to remind you about the meeting. We can also email or text you, but in our experience, these methods are less effective. It's sometimes convenient to "miss" an email or a text, but it's not easy to ignore a phone call from your Productivity Partner.

On WorkMode, you always work with the same Productivity Partner. People with social anxiety appreciate this consistency. It also helps build a relationship with them. They get to know you, understand when, why, and how you procrastinate, recognize potential triggers, and help you avoid them.

On Focusmate, you're responsible for showing up and scheduling your sessions. For many, this becomes just another to-do item on their list—one they may fail to complete because, well, they procrastinate. On WorkMode, you just need to connect with us the first time, and we'll take that responsibility from you. We'll ensure that another session is scheduled, remind you about it, and call you before it starts.

Screen sharing is much safer on WorkMode. Just make sure you don't break any NDAs.

Marginalia Search is an AI-free[1] search engine, mostly because I don't think search is much helped by AI. The sorts of problems that does better than traditional IR-methods based search are things traditional search engines never did particularly well anyway.

https://search.marginalia.nu/ https://git.marginalia.nu/

[1] For whatever AI means these days. Given everything people claim to be AI in their marketing, there's an argument to be made all software is AI.

We make stingray resistant booties! No AI so far, we’re happy enough just to have “multi threading” :)

https://mydragonskin.com/

Holy moly new fear unlocked, is this a real issue beach goers face? Have never heard of it.

Ya, it’s a real issue, but depending on where you are, you may not have as much to fear — stingrays prefer sandy, gentle, warm coastal waters. As one concrete example, SoCal beaches have a lot of stingrays, and in summer days there are often lines of recent sting victims at lifeguard towers.

There are also different species of stingray, and some sting more than others. The ones that bury themselves in the sand when threatened are more likely to get stepped on (and then sting). Other species tend to run away when frightened instead.

Ahh I see, thank you for the details. Your product looks tempting even just for beach shoes ie walking on rocks, coral. I’ve faced problems with existing products that they either have holes in them that small rocks can get in or they’re not tight enough and can easily fall off when swimming.

Ya, it can be hard to balance all the aspects you care about. Unfortunately we don’t recommend wearing our booties on rocks/concrete/coral, because the soles will wear out faster than normal. It’s on the horizon as something to improve. I’d love to give you a recommendation for an alternative, but my personal use case isn’t generally swimming — I haven’t tested a lot of the alternatives for that. Would normal surfing booties work? They’re often nearly skintight, and many have a rubberized sole which should help with the rocks/coral

Eg https://www.patagonia.com/product/r5-yulex-round-toe-winter-...

Wow I had no idea those existed, that’s very helpful thanks!

No problem at all!! I would recommend staying with a “name band”, the generic ones you find on Amazon aren’t nearly as good. Name brand here would be like O’Neill, ripcurl, xcel, Vissla, Patagonia, etc

I am. Learning management platform, with MVP focused on the all-employee security training required by many certifications (Such as ISO27001, which requires repeat training every 6 months. Other regulatory training coming soon, too).

My "product" is to work with companies to create additional company-specific training, such as induction training, product training, etc.

Still didn't build a web page for it yet (I have only two customers), but the demo is:

https://demo.skillful-training.com

Username: one@example.com, two@example.com ... ten@example.com Password: the single digit `1` for all the users

I've been pricing the security training at about half what the rest of the industry charges (pricing is per user, roughly 3USD/user/month)

There's a builk-invite option so that new clients can simply upload CSVs of their users (Email,Name).

I'm building a web extension for an Inuit language (ᐃᓄᒃᑎᑐᑦ; more details in my past comments). It's my first extension, and though it bucks the AI trend, it's still hip and cool because it's in Rust and WebAssembly ;)

To be serious though, it's also using Python and JavaScript, and I'm struggling with slogging through the JS portion. It started out as a mobile keyboard, then I decided to focus on the extension for now. I've also been struggling with connecting with people that would use this extension, so I'm starting to think it won't provide value at all, and I'm considering abandoning it.

I'm in my 3rd year with Notion Backups, helping Notion users back up and restore their workspaces:

https://notionbackups.com

I'm working on a project to find the best places for you to live (in the US).

Right now, people are deciding where to move based on gut feels, spreadsheets, and wikipedia pages. There needs to be a better way to enter personal preferences and come out with enough data that ensures the place you're in is the best fit for you.

Enter your preferences, narrow down locations, and compare them:

https://exoroad.com

Really Cool,had this problem when I was looking to move places. Do you reckon there is a profitable business model with this?

> best fit

Depending on who you ask, and/or by what methods you're computing this, this is in fact AI. Probably most would agree that a neural network is AI, but quite a few will assert that linear regression is too.

edit: didn't mean to make this a put-down or try to exclude you. I guess I'm just being pedantic for no reason (or bitterness about the near-meaninglessness of the term "AI").

That's a fair critique for the context of this thread. There's no linear regression currently. I didn't take the thread title to mean "explicitly attempting to have my product devoid of any AI," as opposed to "building a product, and it doesn't have AI", which is my current case.

Hey, this is really great! I used gut feels, spreadsheets, and wikipedia pages when I decided where I wanted to live. One thing that was really important to me, that your site doesn't cover, is growth. I grew up in a small town, and hated the "we're a small town where everyone knows everyone one and we want to keep it that way". I specifically filtered out cites that didn't show Year over Year growth, which surprisingly filtered out hundreds of cities and virtually every small town in America. You could create a section for "History" and then start with census data and work your way out. The age of the town probably reveals a lot too, like the average age of houses, types of jobs, etc...

Thanks for the feedback! Growth is an interesting concept I hadn't heard people wanted yet, but totally makes sense. Population and home price over time for a specific place, of course. But to filter or weight choices based on YoY growth in some aspect? There's definitely a case to be made for the momentum of some places to keep getting better and others that stagnate.

While I am indeed working on an LLM interface, my other big project at the moment is a tab manager for Firefox which doesn't have a single AI feature, and it will likely never have.

https://addons.mozilla.org/en-US/firefox/addon/grasshopper-u...

https://tonguee.xyz

Free version of duolingo's match madness using top 1000 most frequently used Spanish words. I actually like the competing with others gamification, not so much streaks etc. Made it mostly for myself, as a way to procrastinate instead of actually learning spanish

This looks so great! Do you have any plans to expand to other languages? I'd love something like this for Swedish.

TBH not really, although it would be really easy. I'd open source it if the code would not be an absolute mess

If you do want to open source it someday I'd be happy to help clean up the code in my spare time. I'm actively learning Swedish and would selfishly love something like this :D

Andrew is a web server to help you write a web page like you're back in the 90s. It consumes and renders web pages from the file system, no databases involved. This is the basic design restriction that informs feature decisions. You get started by writing an "index.html" file and then running Andrew from that directory.

It helps with some chores that show up with filesystem based sites, like allowing you to embed a {{ go template }} that'll be replaced with an automatically generated table of contents, automatically rendering a sitemap.xml and a few other things.

It can load a private key and cert for SSL traffic, so it's just a little lightweight tool that does one thing reasonably decently.

I'm just reading the activitypub RFC so I can add support for federated identity. I'm pretty excited.

github.com/playtechnique/andrew

We’re working on a consumer facing self-hosted Minecraft server solution: https://fractalhome.run

That said it’s powered by a general purpose distributed computing architecture so it will probably end up running AI workloads at some point.

Many businesses overlook the significant revenue impact of bugs, even in simple elements like a CTA button. we've developed a tool to help mitigate these issues:

Automate Testing: Save time by automating your testing process. A 10-minute manual test can be done in under a minute.

Prevent Revenue Leakage: Ensure your website's functionality is bug-free to prevent revenue loss.

Timely Updates: Receive real-time alerts with step-by-step video recordings of failed tests to quickly identify and fix bugs.

Here’s how it can benefit your business:

Automated Testing: Ensure all aspects of your website, including CTAs and forms, work seamlessly.

Real-Time Alerts: Get immediate updates with detailed video recordings of failed tests.

Efficient Troubleshooting: Reduce downtime and keep your website functional.

I’ve been working on https://console.wut.dev as an alternative, simpler AWS console. Given that AWS has recently started embedding Q (their AI tool) into their UI, I’ll likely keep Wut AI-free.

That being said, there are a few features I’ve been thinking of where AI could theoretically make sense (like summarizing recent changes to cloud resources from CloudTrail logs) but if I build that, I’m going to focus on the feature/use case and not try to just “jam AI into it.”

I'm building an app where the competition basically is all about AI. It's a "continuous screen recorder" to look back at what you've worked on over the days or weeks. https://ScreenMemory.app

All other similar apps (Rewind.ai is the most notable, or maybe Microsofts Recall) basically focus (focused?) on summarizing your day through AI. Personally I am fine with a smaller scope, instead focusing on the visual aspect of a day.

Most folks here are probably tired of hearing about it, but I work on https://onlineornot.com - uptime monitoring and status pages for software teams.

My aim is to build something better than what's already out there.

I've worked at companies with proactive monitoring like OnlineOrNot before, and was surprised by how bad existing products were. Adding AI isn't going to help them.

I'm building a loyalty points platform.

Started off as a one off solution for a single client, but then I realized that I could generalize it.

The one off solution is running for my client for the last year, but the generalized API is still just a docker container on my PC.

I've been starting to write about the dev process[1]

[1]https://darrelld.com/signal/21/

We're running SimpleBackups (https://simplebackups.com) a backup solution for databases, servers, storage, and applications.

Sounds like a boring business but I love it! What is funny is the number of times we've brought the "AI" buzzword in our discussions: "Maybe we could rename our anomaly detection to AI xYz, it's AI no?" :D

I’m working on a product for a client, and the amount of pressure I’ve been under to “make it AI” has been tremendous.

“Could we add a chat bot?”

“What about AI generated images?”

This is for an extremely regulated industry, where content needs to be approved by dozens of people prior to being used.

(I realise without the context of the actual product this sounds less impressive than it does to me)

I actually tried to shoehorn AI into https://code.movie/ as an experiment once, but it only got unpredictable and of course extremely slow. Hand-tweaked magic numbers plus escape hatches for manual intervention feel much more reliable.

My first Arduino kit ever arrived yesterday, been playing around with it with my siblings.

The rock paper scissors game with an lcd and 3 buttons does not use AI despite it saying

  Player: Rock
  AI: Scissors

That's just so no one realizes getAiPick is just random(3)....

I don't really want to bring much attention to this right now, but: https://github.com/petabyt/fudge Fudge is an alternative to Fujifilms camera pairing app.

Amazing work! It was disappointing that Fuji made the Xapp require location permission.

This is a platform built for readers. It alerts you when your favorite online novels get updated. You can also create lists and review your favorite works (shipping soon).

https://storywatch.org

https://ksylvestre.itch.io/mightygrep

I made a file search utility for the features I've always wanted: search history, bookmarks, and more

DataGridXL, an Excel-like data grid for JavaScript. The manual Excel-like step is not going anywere soon is my estimate. https://datagridxl.com

I’m building a curated blog search engine and reader: https://minifeed.net/

No AI and no plans for any. Drop me a line if you want an invite.

I am building https://www.jobkit.io - a job board platform for your community or niche.

No interest or intentions to touch anything AI.

I'm building a platform to help companies efficiently organize hybrid work, currently without the use of AI. If anyone is interested, we're looking for a CTO ;)

can you elaborate further?

I wonder when we'll start marketing our zero-AI products as "artisanal", "organic", "no GMO", "hallucination-free"...

« Made from 96.538216% natural ingredients »

"Made from 100% free-trade, handcrafted paperclips"

"Hallucination-free" gets my vote!

Everything we were building was AI free. Now everything has AI on the roadmap.

slack & zoom. all in one place. http://teracy.io

Working on a new free and open source social platform called Kowloon that doesn't even have "algorithms" (unless you count sorting posts by date as an algorithm), much less AI.

When I was a kid I went to the Media Lab and got to meet a few of the pioneers and got my copy of Society Of Mind from Dr. Minsky's wife Gloria. I wanted to build AI so bad.

Now I'm middle-aged and have zero desire to dick around with Markov chains on meth so marketing scum too lazy or stupid to even write their own copy can pimp their crappy products that no one wants or needs to a dying planet. I'd rather go build furniture on an Amish farm.

In lieu of that, the only thing I use AI for is as an autocomplete in VS Code.

Yeah, building VisualJS, an AI free coding tool:

https://github.com/yazz/VisualJS

Does blackbox optimization count as AI? Or is it only LLMs at this point?

Sounds interesting. Do yo mind sharing more details?

Reply on news.ycombinator.com