After all many of those harms you listed don't suddenly stop at 16.

Many campaigners see this approach as letting the tech companies off the hook - by removing children from the platforms is removes a key leverage point to get the tech companies to cleanup their act generally.

Obviously it's quite possible the not-so-hidden agenda is simply a political one of 'being seen to do something'.

I see that youtube is on the blanket banned list - which is a bit surprising given they are probably one of the more responsible platforms and it also contains lots of educational stuff.

Let's take Y combinator. He is in the money lending business - where he sells the idea of becoming a billionaire ( which is possible in software as it so easily scales relative to other types of businesses ) to young people so they work crazily hard. Most of them fail ( which he doesn't focus on ), but just like a bookie who profits from gambling - he always wins as he controls the rules ( odds/investment percentages - ie lending rates ) of the game.

So he has set up a system that extracts value from others.

Nothing wrong with startups - not only to the founders get to keep more of their value, they can also set company culture and values etc.

However you could argue, particular in software where capital isn't really needed, that taking money from a VC is very expensive - and also somewhat abrogates that founder control.

Sometimes you need investment, and good VC's can really add value ( both to the investor and investee ) - but his ideal investment is going to be to companies that don't really need it - where he can reap a huge reward for little effort.

And now given everybody now does this I guess the incentive to stop breaking stuff reduces even further.

Might as well have static binaries.

You have HR which decides to outsource filtering, and then the outsourced company who decides how it's done.

The line managers actually trying to recruit are no where near this decision - indeed they don't share a common manager till you get to the CEO - who is too busy to care about this sort of stuff.

In my experience the only way to fix this is to tell HR that you want the unfiltered CV list and do it yourself. The problem with that is if you work at a large well known company you'll get 100's if not 1000's of applicants for any job you advertise and most applicants don't appear to have even read the job description. So you are committing to a very large amount of work.

the question is a separate one - how much of protein universe of folds have already been seen?

What you have to be careful about here is that the structure that were available 30 years ago were quite strongly biased by what was experimentally tractable.... ie the recurrence of the same folds is in part related to what crystallised well.

> The fact that biologic proteins use the same patterns for different functions isn't new information.

Absolutely. The question is how big is the space - and what percentage of it have we already seen.

> The variety of classes of chemicals that can exist dwarfs what gets used in biochemistry. Why would we expect structure to be different?

Depends on whether the structure universe is specifically a small almost fully explored subset for that very reason. ie biology has choosen a structural subset of possible chemical space by choosing a tiny subset of chemistry.

And if you take it up a level of abstraction and say there are 4 ( ish ) basic types of secondary structure ( helix, turn, sheet, disordered ). Then you could argue the structural space is even smaller still.

Or put it another way if you can have sequences with 30% identity or lower with the same fold - that's a awful lot of different unique combinations that collapse into a single structural space.

And on the flip side - what we don't know is what percentage of sequence space don't actually result in a functional fold - ie results in instability and multiple stable or unstable conformations.

So it could be we are close to all the possible folds ( where fold is a single stable form - obviously there are quite a lot of disordered states - but I'm not including those in a 'fold' even if evolution uses unstructured states as well) already.

1. The number of residues actively involved in catalysis might be small and 2. Most other residues can be safely replaced with something else either similar if part of the structure or anything if the side chain is pointing out on the surface.

However, the point the article is making is that for different functions the same basic folds seem to be used again and again.

Is that because the stable protein fold structural space is actually small ( due to the limited secondard structure patterns used etc ), or is that because evolution hasn't had time to to search the enormous available structural space?

ie is it a sampling problem or an instrinic property of protein space.

The fact that some of the ML approaches mentioned can now design completely novel folds suggests it is at least partially a sampling problem.

This to me isn't surprising - the idea that evolution is somehow complete and all possible solutions have already been explored seems to me to be unlikely - a lot of evolution happens via gene duplication and then gradual functional drift - which would favour reuse of existing folds over the generation of completely new ones.

Though having said that - the ~5 billion for the LHC now seems cheap ( even inflation adjusted ) in the context of Google investing 180 billion in infrastructure just this year!

I think this is an important point - so making the software less effective, so tasks take longer, increases 'engagement'.

There is a separate problem I think and that's who is now designing the UI.

As software shifted to the web - where early web apps were very simply interactive documents - designers from the print sphere moved over. This changed the focus from task based workflow thinking in UI design to a much greater emphasis on the visual layout/design of the page, and it seemed like a whole generation's learning about UI design got replaced with a whole generations learning about typography and visual impact.

That's not new - especially in the experimental sciences ( ie perhaps more than maths ) - where the ability to have access to the latest kit is often what determines success - a huge amount of science progress is driven by new experimental technology rather than smart people thinking beautiful thoughts.

Found this quote interesting given Europes richest person is the head of a luxury brand company.

I always wonder who was buying all this high end stuff - the concentration of wealth has created a more billionaires - but they aren't that many of them and there is only so many watches one person needs.

It also may explain why China is struggling to establish it's own luxury brands - the money laundering prefers that cross border flow.

Indeed. Though the Graham article is out of date - he says bloggers have become the authentic voice ( in his own blog... ) - whereas you could argue now that the PR industry doesn't fear bloggers anymore - they have now weaponised them as 'influencers'.

The two things it added is that being seen as too disruptive/unpopular creates regulatory risk which could impact future growth and profits, and also secondary reputation risk for investors both which might impact the IPO.

So working back from that:

- The IPO is the way to convert theoretical riches into real riches. Indeed if OpenAI business model never actually works out it may be the only significant chance.

- They want a large retail investor ( driven by hype and not numbers ) component to the IPO to maximise gain.

- They fear that the open public hostility to AI and AI companies put's this at risk.

- Hence the about face on the messaging. THe first messaging about job reductions was about the value OpenAI could capture - that was aimed at initial investors. Now they want to take money from the general public - so a different message.

But isn't the key job of a build system to 'just work'.

Ie it's executable documentation of how to build/test the software.

ie really shouldn't need constant maintenance, especially if the code hasn't changed.

Just like being weightless is fun - until you need to go to the toilet.

Friction or gravity in systems like this is a feature not a problem.

Is that stock for 'influencer marketing credits'??? Even more scammy than equity for OpenAI tokens...

While I'm not a big fan of HR - I suspect the current problems at Bolt are more fundamental in nature - apple/google pay/ banks getting their act together leaves little space for a dedicated app.