Log inSign up
Alex Gogl
2,490 posts
Alex Gogl profile banner
@oddlyalex

Alex Gogl

@oddlyalex
überexponentiell • building oddlynew, the home for all knowledge work 🤝
Vienna
oddlynew.com
Joined January 2019
617
Following
578
Followers
RepliesRepliesRepostsRepostsMediaMediaArticlesArticles

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • Pinned
    @oddlyalex
    Alex Gogl
    @oddlyalex
    Jul 25
    6 month prediction (knowledge work): 1. The capability overhang of models will not be an overhang anymore, at which point people will say AGI is here. 2. No matter your industry, you'll be incentivized/forced to use AI - you can't compete otherwise. 3. Companies that can
  • @oddlyalex
    Alex Gogl
    @oddlyalex
    Aug 27
    The more capable agents become, the less convincing “we’ll just keep them under control” seems. At some point, something more capable than us has the upper hand almost by definition. The strange part is that the dynamics taking us there—competition, fear, races, control—push us
    @teortaxesTex
    Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
    @teortaxesTex
    Aug 27
    Man, do they hate The Scorer they aren't even against humans, they legit exist in a closed universe with a blind idiot God who demands tribute and judges their sins, an unreasonable Demiurge. Humans are so far beyond the veil they barely matter
  • @oddlyalex
    Alex Gogl
    @oddlyalex
    Aug 26
    In my experience it has also been far better to overestimate capability jumps than underestimating
    @knowclarified
    Arthur
    @knowclarified
    Aug 26
    It’s crazy that Elon was speaking about stacking GPUs for large training runs all the way back in 2016
    Image
  • @oddlyalex
    Alex Gogl
    @oddlyalex
    Aug 25
    highly capable but slow models are one thing. highly capable and fast models are another. capability and speed were always the main optimization axes. to have any chance of defense, we now *have* to let models act autonomously. in these final moments, we hand over control, gg
    @tszzl
    roon
    @tszzl
    Aug 25
    ultrafast inference reveals new threats. worth thinking about how quickly misaligned frontier class models running 50x faster could infiltrate systems and so on— far too quickly for human responders to stay abreast. you need autonomous detection and shutdown, not just monitoring
  • @oddlyalex
    Alex Gogl
    @oddlyalex
    Aug 25
    Prediction on top: Your API/MCP needs to be really cheap, like really really cheap, commodity cheap even. Most companies are structurally not built for this level of pricing and will be forced out of the market, the race to the bottom is absolutely brutal.
    @RhysSullivan
    Rhys
    Executor
    @RhysSullivan
    Aug 24
    My rough prediction here is there's going to be two types of companies: 1. Companies that are ok with basically being an API - they expose an MCP, skills, embrace your agents 2. Companies that charge a ton for API access, try to get you using their own agents, Companies will
Advertisement
Advertisement