shisisms

There have been a range from quite basic things like social media posting, that were a real pain to organise, but actually took no time at all - to external APIs around text summarisation or modification, right the way through to modifying images etc. - the external APIs had a range of different benefits, but always were priced annoyingly enough that it was faster to just build what was needed.

Under enable early access apps, but rolling out over next 15 days apparently

Totally agree. And beyond that human life assumes it’s own sentience through an incredibly narrow view. Isn’t it fair/logical to assume that if giant mushrooms did over the earth that given the passing of considerable time they’ve likely evolved to a far more considerable state of intelligence/harmony than humans have. Their survivability/adaptability being an illustration.

Likewise, one ends up sounding like an anti climate nut, but the existential fear for the planet seems arrogant at best and massively over emphasises our place in the universe. See, I sound like an anti climate nut.

This is beautifully executed, well done. The UI in particular is near flawless for what it needs to do.

Hopefully you add other languages soon? Are each of the lessons static or are they dynamically generated?

Concept is great. AI to full page would be even better. And a couple of links to demo examples of full sites would be a deal closer. Requires too much effort to think through at the moment.

All the best!

Here we go, ding ding, the scrap begins…

Yes, there’s a really sensible approach here. A recent experiment saw a fairly sophisticated internal tool use a number of sheets as a backend. Aside from speed of implementation the other real benefit was the ease of which non-technical users could not only make easy edits to the backend, but also that they could onward integrate to other APIs/data sources through sheets.

There’s an irony in the fact that the more the slips that are covered by the press and the Twitter bubble chamber -the more desensitised everyones’s getting to them - no matter how outlandish, offensive or inappropriate. Leading to a more dangerous position where there’s a growing acceptance of the obvious fallibility of current LLMs.

Maybe Google is right to be cautious. In fact, they almost certainly are, I hope they hold strong for a while longer at least.

Love the point around the iPhone in particular. Part of the joy of the object is using it naked, feels as it should and what a joy that is.

Putting a beautiful work of art in bubble wrapping May we’ll protect it, but would diminish its joy, beauty and sheer pleasure to that of bubble wrap.

Not being a software engineer I’ve been relearning to code again, and after a deep dive dived into a new framework. Which was a headf*k to put it mildly including server config etc.

What I re-learned was the “arc of despair” that initial period of complete disaster where everything wasn’t just confusing but incredibly difficult. And many many times of asking myself of whether this was worth it and whether I shouldn’t just pay a “professional” to do it. Then within days I was swimming fluently. Was a strong reminder of the various stages of being auto-didactic. The need for faith :)

And yes, in the end it was worth it, because the lateral connections and first principles you learn (complete with new learnings and questions) allows you to go much deeper than simply paying someone else to do it. Was a reasonably profound experience.

This is just really good fun.

It was a few years ago when Google got absolutely panned for claiming they could call a restaurant up and book a table using an LLM + voice. And here we are where it now feels not just acceptable but fun.

Wild ride ahead.

Valiant, but isn’t the entire point that the cost of generation is so low you don’t need a curator.

Any of the APIs can just spit out 100 variants for far less than $5, so why bother with curation?

Nope, that’s impressive and not dissimilar to the EMM: https://emm.newsbrief.eu/

I’m thinking more extrapolation of news eg https://www.bbc.co.uk/news/live/uk-politics-63327087 so not focussed on multiple events, but rather singular events, with more frequent updates.

So starting point is identify key trending global news event eg major terrorist attack, election, major domestic incident, maybe even pop cultural. And then rather than relying on disparate news sources to update, bring them all into one place.

Yep, that’s sensible, if not easy. What about rumours? Is there value in adding news that the major publishers can’t, given the diligence they must apply, but here could be brought in, albeit with a clear UI that makes it clear it’s unverified.

This is incredibly well done! Congrats. It’s such a solid layer on top of recent developments and offers instant value.

Wish you much success!!

That was a good skim. Already seeing wafer thin, and excellent, consumer apps built on stable diffusion.

I’ve been increasingly concerned that given now the wide spread skills to build consumer apps for example, the value will be eaten up too quickly.

Thai article makes the point that in fact this isn’t where (economic) value really sits. Building a moat requires more than just the ability to build.

Thus, it feels as though building simple consumer interfaces to the latest AI models is a short-term and largely thankless play.

Delighted this is in Facebook’s hands. No doubt responsible and careful thought of implications have been baked in from the start.

Agreed, but what if it wasn’t. Google, non-gamed, in the earliest days implicitly surfaced trust. That’s now lost. Reddit has the same problem, it doesn’t surface trust (karma isn’t the same thing).

Real life communities have trust built in through a range of mechanisms. That’s missing from search today. Which is trying to rapidly answer a question, rather than deliver a trusted response.

Standard response, but it’s not true is it. Again, level of trust in the person posting is minimally measured, so again the cognitive load is placed on the searcher. Which means that at scale it devolves into nonsense.

Google is the “wrong trousers” maybe the almost perfect translation.