westoque

This weekend I’m working on mobile app that allows you to upload photos and it turns everything into a stitched anime (ghibli or not) with movements and eventually sound and script. Great to make mini animes from your day or travels or anything.

not to hijack this thread but my dad did extensive research in sea breeze and rainfall modeling and he would have loved to see these AI and machine learning advancements in weather prediction.

[0]: https://www.revistascca.unam.mx/atm/index.php/atm/article/do...

[1]: https://wmo.int/about-wmo/awards/international-meteorologica...

working on cursor for desktop. why rely on AI agent that’s self-contained when it’s limited, can’t access the browser, can’t open apps or click around.

i simply want mine to be able to fill in forms in preview with a passport image as context. also to be able to do recurring tasks as if i was the desktop user. e.g., i’m going to bed keep working on this spreadsheet.

it’s working and built but very slow and buggy atm. uses multimodal LLMS and OCR but lots more optimizations needed. need to make it a lot faster. can demo it and need help if anyone is interested.

i've worked on many legacy code bases and i think an underrated skill that you need is courage.

courage to be able to push changes you think could be breaking changes but always be optimistic that you to be able to fix as soon as possible. tests are ideal, but not everyone has 100% coverage. sometimes you just have to trust your gut and do diligence and push the change. don't be scared to push something because you don't know the behavior in production. the only way to know is to try.

reminds me of the anthropic claude jailbreak challenge which only pays around $10,000. if you drive the price up, i'm pretty sure you'll get some takers. incentives are not aligned.

I'm a CTO and scaled multiple companies from zero to multiples of TBs of data (also to millions in revenue). I also worked on companies such as yourself that have an existing solution and want to revamp their entire technology base.

The short answer is yes, you should hire an in-house team so you can execute faster and also reduce the red tape of the provider not being more proactive.

How to do this is more nuanced:

1. I would start by documenting and/or asking for documentation on how the whole system works. This is key. I would start with a high level diagram. Even if you're not technical, this would allow you to evaluate the existing stack and see what could be worked on in parallel to the existing tech if any. I feel like you are technical enough to understand this since you understand CRUD and other terminologies. You can even set a time for the provider to explain the services to you if needed. I say parallel because I'm assuming you can't just quit using the system entirely, it will need to function until you do the switchover.

2. Once you have a good understanding of the system, this will allow you to be more specific on your needs. I almost never recommend a re-write but based on the statement that you are using a third-party product, this might be the case here. If it is indeed a re-write, I would approach it so that it's 1-1 parity with the existing system first, and if not, maybe with some ample planning and product grooming make the features even better. Furthermore, this will allow you to make a basic PRD for the requirements of what you are building. Doesn't have to be really specific but it will make it clear to yourself and others on what the task would be.

3. Hiring. At this point, I would start hiring a CTO/VP/senior engineer. Describe the problem from 2., maybe deep dive even if needed and start to get technical leadership going. I'm assuming cash is not a problem so if you start at $175k/year salary (maybe some stock/benefits) you can find many technical leaders that would go for this. Once you make that first hire, if they have a good network and or social clout it should be easy for them to hire people and/or consultants to do the job.

4. Iterate. Once you have that amazing team and if they have a good process in place they should keep iterating on the product and continue doing this on a steady pace.

5. Once the product is ready or any point that is it deployable make the switchover. At this point, you should have full in-house control of the product and have competent tech team in hand to do whatever is needed.

NVDA market cap for sure is bonkers but AI is here to stay. Anytime you question the future, think if you can see yourself in "that" future. For me, AI is already a part of my workflow and it's beneficial to my daily life. From coding, explainers, and to random questions in perplexity.

In coding, it made me a lot more productive. No more do I have to scour Google or Stack Overflow. Instead, I get the answer straight from the AI. It might not be working at times but it's definitely better.

Even random questions like "What is the cost to produce the Apple Pro Display XDR?" is being answered more promptly by AI.

This is contrary to the crypto craze where there was expected use but sadly it did not come to fruition. And until now, it's still a speculative industry, a solution looking for a problem.

just basing my assumptions on:

- https://www.theregister.com/2024/02/08/apple_web_apps_eu/

- https://itnext.io/apple-wants-to-kill-pwas-0895be2e497b

i really hope they don't as it does provide an acceptable alternative to native apps. not everyone has a mobile team that can build native. imagine a small team just making a web app could now also provide a similar experience and value for mobile out of the box.

alternatively you can get away from the app size bloat by using X as a PWA (progressive web app). in iOS, this would use safari as the app container and you would just have the assets to download for dependencies. it's sad however that apple is now trying to kill PWAs. i wish people would use PWAs more as this actually makes it easier to build apps.

Most of the speed I believe is from using the server iodine https://github.com/boazsegev/iodine which is a wrapper around facil.io https://facil.io that is built using C.

Great project either way!

I was using the Kinesis Advantage 2 (KA2) for the longest while 8+ years now, and then recently upgraded to the Kinesis Advantage 360 and the KA2 was still hands down the best keyboard that I used. I still switch between the KA2 and KA360 but the KA360 is as mentioned plagued with so many issues, from the bluetooth connection issues all the way to setting the keymap using the ZMK software. I would still recommend the KA2 for most users and also the best bang for the buck if you can get one used or refurbished. The split also doesn't make as big of a difference as it seemed at least from my proportions.

regarding concurrency, language plays an important role and pretty much dictates how code is written which is i believe where it has most of the frustration. in erlang it’s in a functional style, in javascript it’s in an asynchronous style. what i’ve come to realize is that it’s still better and more maintainable to think synchronously and have the core of the tech handle concurrency, for example golang with it’s goroutines or multi process in ruby/python. admittedly it’s not as concurrent/distributed as erlang but should be easier to work with.

another similar concept i have found to be effective is to “teach what you learn”, i have noticed that it takes some good understanding of the topic you want to teach in order to communicate that effectively as well.

I've seen this in person. It's such a beautiful piece! Note the details in the background and the 3D perspective of the frame. Easy to miss but when you focus, you see great thought and detail that goes into the piece.

same with high end luxury watches. i think that these companies just add another layer on top of the standard veblen goods definition making it even more exclusive/desirable, and therefore even harder to get one.

this.

properly made abstractions are actually more maintainable and easier to manage. for ex, take anybody learning programming about writing to a file, if you write your code using ruby, it would be a lot safer and maintainable than the C counterpart even though the latter is more powerful.

As a software engineer, I wonder how accurate they stuck to their said schedule at the last page.

I'm a fan of RSpec and any BDD (Behavioral Driven Development) type of framework. You are correct that they all do the same thing, but for me everything boils down to which one provides the quickest understanding of the code by the reader.

There's another testing library in Ruby that is called minitest. The syntax is different, for example you would write "assert_equal num1, numb2". My problem with this is that you don't know which is the expected vs the actual values. Is it the first or second argument? On the otherhand, my framework of choice, RSpec would write this as "expect(num1).to eq(numb2)". The syntactic sugar alone makes this easier to understand.

> When having conversations in HN does it bother you that you have no way to know if you're talking to a real human, not hiding behind fake usernames?

On HN, I'm not bothered at all. If I see the username "dang" for example, I know it's YC moderator dang also same with "patio11". I see HN as more of a private group discourse vs Twitter where it's public and where blue makes more sense. Akin to a forum, where I frequent like home-barista, the community is smaller and more focused so I don't see why having blue makes sense.

Funny. I see it differently. I pay for the blue check because it adds legitimacy. If I'm replying and having conversations with people with similar blue checks then I know I'm talking to real humans not hiding behind fake usernames. Also, any threats or anything from bad actors could otherwise be tracked based on payment method.

When Teslas were hot, it was really hard to buy one in my area, it involved using the search function on the Tesla website, inputting my zipcode and waiting every few hours until new cars are out. I basically made a script that texts me if there’s new car with the configuration I want with a link to the buy page so I can buy it immediately.

random but this is probably a good use of blockchain tech. a smart contract that stores the notes and unlocks using a key that is given to a loved one. perfect for use case for a lasting note. only time this goes down is if the entire network goes down.

i actually support the pay to verify feature. my reasoning for this is it gives authenticity to the user. if the user is verified then i know that the user is most likely not a bot account and gives some credibility. could this be free? certainly, it’s how the old twitter verified works but now since all users can be verified, it democratizes this feature and i think its better overall. i notced discussions with similarly verified users are more civilized probably due to the personal information being shared and makes them think twice about doing anything malicious or otherwise.

I was CTO of a company that was doing consumer text messaging through Twilio. I would say Twilio actually does a good job of monitoring messages going through its system. Our system at times were used by spammers to send spam text and Twilio's automated bot would catch these and would then ask us for reports if we have permission to send automated text messages to the numbers.

Same here. I am a big proponent of state machines. I have found that state machines bring sanity to an otherwise complex situation. For example, most inexperienced developers will check presence of certain fields to imply where someone is in the process whereas it would be better written using a state machine.

I am this exact position. I had a separate account (work related) and personal one. It boiled down to me knowing that I value influence and authenticity more so I used my personal account. It's easier to make arguments and discussions with people knowing who you are vs a pseudonym. You do need to be a bit careful about words (as you always should) because people interpret them differently. Also, accept the fact that you cannot please everyone and that there will be people that are for and against you.

If you find the message above useful, please do follow me on Twitter: https://twitter.com/westoque

this is one of the reasons i changed my OS to linux since docker there works natively. i believe there will always be a bottleneck for peak performance on macOS until we are able to run natively, not virtualized.

> I don't care about money

I’ve heard this statement used frequently in business. I think this statement is incomplete… what if instead you said something like “I don’t care about money as long as you get it done before the deadline”, wouldn’t that sound as reasonable? Think of money as a tool, when used responsibly could do good things. I see a lot of people being cuffed not spend money just because it sounds “expensive” but really if you 2x’d that spend then it’s money well spent.

Thank you :)

I've been a CTO for 5+ years and have grown a company from 1 to 15+ engineers.

I agree that a CTO SHOULD be technical but it shouldn't be their only skill. The CTO's skill actually varies depending on the stage of the company and being technical, helps in making decisions in every stage. A core part of being a CTO is to set the technical direction of the company, you are a Chief Technology Officer after all. A part of a CTOs job is to know how to execute/deliver the plan of the company in the most efficient way technologically speaking. As the leader in the engineering department, if you have say a budget of 1 million USD, not only do you have to know how to allocate this efficiently, but also know the technical details of why you do so in the grander scheme of things of building the product.

As an example of my said stages, if you are with a company that has:

1-5 engineers: The CTO is mostly likely coding and also making decisions on what technology stack to use and sets basic processes for development (kanban/scrum or basic todo list).

5-15 engineers: The CTO is mostly doing process management and technical analysis of the stack to set further systems and processes for future growth. Here you will start to find the issues of the technology, you are growing the engineering team because there is a further NEED to hire. For example, you hire more frontend engineers because the plan is to build a more ambitious frontend.

15+ engineers: Here you already have established good processes and have the product moving forward with teams being mostly self sufficient. The CTOs job at this point is to work more with product to know the future roadmap and efficiently plan further hires or changes technically speaking.

My few c's from this. I've yet to experience growing a company for 50+ engineers but that is a goal that I have still going forward.