A point of confusion is where it says both:

》Want unlimited launches of the latest launchpads? Need ongoing support? Reach out to (email)

》Reuse your code as many times as you like.

Is the first one a cloud offering and second referring to self hosted?

Thanks for flagging that! The intent is that you're welcome (and encouraged!) to use your launchpad as many times as you like, for as many projects as you like.

However, that code will slowly become out-of-date in the form of deprecations, dependency updates, new launchpad features, and a changing tech landscape. If you always want the most up-to-date starting point, that requires another (discounted) purchase, or the ongoing support plan!

Does that make sense?

It does thanks. Sounds like "support plus access to future versions when released"

A pricing table might help here.

I think you're right. I super appreciate the feedback!

You need a lot more power. I found gpt4o struggles doing basic OCR of printed text by hallucinating alot, while tesseract engine (old skool) gets it perfect. You need the model to be powerful enough to do everything.

You can work around this by the way by sending the output through a checking stage.

So picture -> gpt4o -> out1, picture -> tesseract -> out2, out1,out2 -> llm.

Might work for sound too.

Interesting, I've actually been using gpt4o extensively for OCR and didn't encounter any significant issues - could I ask you to please give an example of an image of (otherwise legible) text that it hallucinates on?

I'll send you an email.

Same, it's perfect at OCR. Generating an image with text in it however… nope!

Speech is inherently easier to represent as a sequence of tokens than a high-resolution image.

Best speech to text is already NN transformer based anyway, so in theory it's only better to use a combined model

The built in DB makes this very interesting to me. Although I would prefer sqlite. The low LOC count is nice to know.

Minimilism is something that is hard to get right on the web. The cliche is people avoid django because it is too complex then hand build all its features anyway over time!

However some jobs are minimal esp. internal dashboards so there is a place for it. You may find fasthtml which is also minimal interesting.

Curious: why Java and not Go? Go seems more in line with being minimal. Especially for deployment.

And that day, I was told I passed the Amazon interview and was offered a job as a L4 Software Engineer.

I believe LLMs decimating the role of a software engineer requires AGI, which the second that happens decimates all jobs.

What it may do is change the job requrements. Web/JS has decimated (reduced by 90% or more) MFC C++ jobs after all.

The programmer doesnt just write Python. That is the how... not the what.

For clarity: do you want to be a software engineer, platform engineer, data scientist, ML expert or something else?

I asked it to prove Fermats Last Theorem and it started writing (some nonsense) into the margin. Which was funnny.

Cool project tho

Unlike Fermat, the whiteboard has enough room in the margin :)

If you can design a new encoding scheme, zero bits is enough.

Wow. Bike shedding really is expensive.

Related-ish: https://en.wikipedia.org/wiki/Nosebleed_section