Without qualifying, "consensus" is rather ambiguous. Wikip on consensus:
"The level of agreement necessary to finalize a decision is known as a decision rule.[3][7] Possible decision rules for consensus vary within the following range:

    Unanimous agreement
    Unanimous consent (See agreement vs consent below)
    Unanimous agreement minus one vote or two votes
    Unanimous consent minus one vote or two votes
    Super majority thresholds (90%, 80%, 75%, two-thirds, and 60% are common).
    Simple majority
    Executive committee decides
    Person-in-charge decides"

Most of the bickering could be avoided if people define the terms they use/stop conflating consensus (as used within the Bitcoin protocol) with level of consensus required to make changes to the protocol. With so many coders here, kinda' weird that this needs to be brought up.

Lets take all the personalities out of the equation for a moment and look at the stark facts.

What's happening now is that the transaction cost is being driven up - whether intentionally or unintentionally. That drives up the marginal revenue on any proprietary scaling technology. (I don't care if it's open source or not, it's still effectively proprietary since any off-chain scaling infrastructure will be run by centralised third parties almost by definition, just as mining pools are otherwise there wouldn't be any point in them).

If you have large investors with tens of millions of dollars ploughed into such technology who simultaneously employ personel who are at the heart of the influencing the debate over whether bitcoin should be made to scale natively or not, how is that not a conflict of interest ?

To me it's about as glaring, unhealthy a conflict of interest as you could get. As I say, take the personalities out of it - all of whom we must still assume are acting in good faith - and those are the simple facts.

Why can we not have a natively scaling blockchain ? Is it really too much to ask over a decade into the 21st century ?

Bitcoin as a clearing layer for customer facing, proprietary transaction interfaces is no different from JP Morgan with a blockchain in the back room. More money will be made from providing access to the blockchain than will be saved from using it. It may still be big and it may still make money but it won't be free of counterparties anymore. They'll all just creep right back into the equation like they're starting to do now with this blocksize weakness.

Thin end of the wedge.

Give us 32 MB back then  . You could be compatible with the original Satoshi client if you just patched technical mistake with lock openDB making many transactions in block unable to validate. But the consensus was 32 MB block size limit from first day, so obviously this is not such hard consensus rule set in stone to change because it was changed once already...

And Gavin proved to be wrong.  We are still far from hitting the limit for normal transactions, even during peaks.  No block size, not even Gavin-sized blocks, can withstand malicious spam attacks.  Gavin doesn't have a solution for that.  Only the current Bitcoin developers are working to solve the problem, and provide real scalability to Bitcoin.

Don't be a fool.  BIP 109 is a plan to make matters worse through a hostile takeover.  No blocksize can prevent spam attacks.  Gavin hasn't contributed meaningfully to Bitcoin for years.  He is busy collecting money from Coinbase to create diversion and prevent privacy improvements to Bitcoin.

If the goal was to increase capacity, why not reduce the transaction size instead?  It can be halved by using different addresses and signatures.  This is a hard fork as well, and the deployment will be equally hard.  But it is a harder sell, because it isn't as intuitive as a block size increase to the uneducated masses who think larger blocks will prevent spam.  To take control over Bitcoin, Coinbase needs a solution which is easier to sell to the uneducated masses.

Did you try to send btc, with proper fees, before opening this thread?

This backlog seems to be low fee based, therefore not really anything to do with block size.
Block size may be a pressing issue. But this is not why, as explained by Lauda and others, and now me.

Boah, the Gox Sturle and the Staff Girl. Dream Team.

You don't understand the difference between a hard and a soft fork.  A soft fork adds new rules, which can be done without breaking the consensus defined by Bitcoin Core (all versions).

RBF wasnt used, just a straight up double spend with a higher fee (777 satoshi per byte).

btw, I just did a TX, sucked it up and paid the fee core estimated (44 satoshi per byte, big TX, so 60 cent EUR total) and it was confirmed after ~2 minutes.

Yes, Core defines consensus for Bitcoin, and the core consensus rules were set in stone from the point when Satoshi released version 0.1.  If you don't believe me, read his words for it here.  If you see two posts later down, you will see Gavin defining his future role in Bitcoin.

Attempts to break the consensus rules will make an altcoin.  In principle anyone can do a soft fork by enforcing new rules, stricter than the rules in Core, and this will make a soft fork.  It will not succeed however, unless there is a clear majority supporting it.  Older nodes, all versions, will continue working after a soft fork.

I have been using Bitcoin since 2010, and exchange bitcoins for a living.  The health and development of Bitcoin is vital to my business, and for that reason I follow it closely.  I have no connection with the author of the article.

I suspect the main "problem" is that people use fixed fees instead of good estimates (like e.g. via core, electrum or "cointap" provides). At least some of the problematic TX come from services that are most likely based on API calls with fixed fees of 10k satoshi per start kbyte.