Comment

Technical people have been through many hype cycles. Cloud, machine leaning, big data, blockchain, and now ‘AI’. We are just tired of marketing people lying and overhyping technology.

The cloud did not save us money, it costs more to run in Azure, AWS, GCP.
Big Data was social media companies trying to justify sucking up every scrap of data they can.
Machine learning was trying to make algorithms appear smarter at spotting trends and being more reactive.
Blockchain has turned into a massive scam platform with crypto currency and nfts. Any regular database with an Audit table will outperform them. ‘AI’ is shaping up to be a justification for lot of job losses for little to no gain as shown by articles from Gardner and The Register

Coding tools by ‘AI’ companies are reintroducing flaws that we have solved because they are trained on Stack Overflow questions not answers.
Creative ‘AI’ tools do a bad job at producing acceptable output without hyper specific instructions to not have an extra arm or finger.
AI crawlers are slamming every webpage online in the hopes to update models with the latest data even though the data is unchanging.

source

Sort:hotnew top

deegeese@sopuli.xyz ⁨1⁩ ⁨day⁩ ago
Hype cycles come and go, but cloud compute is still around because it has real advantages, and big data solves real scaling problems.

You left out the biggest hype cycle. The dot-com boom wasted a ton of cash but also created Google Amazon and Netflix.

source
- lime@feddit.nu ⁨1⁩ ⁨day⁩ ago
  it has real advantages for some scenarios. finding those scenarios (and by extension, everything it’s not good for) is usually what makes tech bubbles burst.
  
  source
  - deegeese@sopuli.xyz ⁨1⁩ ⁨day⁩ ago
    I miss having a big private cluster.
    
    I do not miss fighting with other departments for a share of that one cluster.
    
    I definitely do not miss waiting months for capacity to be added.
    
    That it costs my employer a bit more is a decision made above my pay grade.
    
    source
    lime@feddit.nu ⁨1⁩ ⁨day⁩ ago
    meanwhile we’re running our entire infra and support for hundreds of international customers on one of those old flat lenovo thinkstations.
    
    we only just got a vps, because of routing issues.
    
    source
    -> View More Comments
- slazer2au@lemmy.world ⁨1⁩ ⁨day⁩ ago
  
  The dot-com boom
  
  In fairness I was in school for those so I have no experience in them. While I was in the workforce for the others.
  
  I am curious what advantages the cloud offers that a colo in a DC with proper planning can’t.
  
  source
  - deegeese@sopuli.xyz ⁨1⁩ ⁨day⁩ ago
    The advantage is you don’t need to plan and provision your own infrastructure in a DC because the vendor handles it.
    
    That flexibility is worth serious money.
    
    source
    slazer2au@lemmy.world ⁨1⁩ ⁨day⁩ ago
    Personally I don’t call that an advantage. Paying 3-5x times in monthly fees over the course of 5 years just so someone else can lifecycle the hardware. That doesn’t even include the performance hit from using shared infrastructure compared to dedicated infrastructure.
    
    Look at 37Signals saving $7 million over 5 years after leaving AWS.
    
    source