The big, big question facing everyone is how do you make an original comment about the AI bubble.

Ian and George tack away from Cisco vs Nvidia and nonsense PE comparisons and revisit ‘F’d Companies spectacular dot-com flameouts’ - the basis for Ian’s recent blog, looking for real lessons and how not to avoid post-bubble mistakes.

Key points are:

Dotcom was lots of relatively small things and a bit random. Much of the headline grabbing nonsense was for ventures with only single digit $m of funding. AI is big money spend in a relatively few areas.

Much of the nonsense worked eventually – webvan, Pets.com and furniture.com – but maybe not Ocado. Beware of throwing away things that don’t work first time around.

Key parallels – Application Service Providers, short term fail but long-term success (as it morphed into SaaS) – will AgenticAI follow this path when it’s just too difficult at the first try?