Investor Jason suggested Perplexity's automatic routing as a fix
@EricNewcomer use @perplexity_ai and it will route automatically -- does a solid job at it. cc @AravSrinivas for background
as a consumer it's getting a little exhausting figuring out what model I need for every particular task
Users are frustrated by the need to mentally map tasks to competing AI models because it feels like the worst part of the stack and they want one model that does everything decently, while some appreciate recommendations like Perplexity AI.
@Jason @EricNewcomer @perplexity_ai @AravSrinivas I agree! @perplexity_ai for the win!
@EricNewcomer The Question is rather GPT 5.5 - xhigh or Fable 5 medium
But within Anthropic - probably Opus 4.8. It´s still early, but so far it seems like Fable 5 is burning an exorbitant amount of tokens
@EricNewcomer Maybe stick to your brain and don’t outsource your thinking to some model. Less confusing.
@EricNewcomer You’re going to be seeing a lot more things like this.
https://factory.ai/news/factory-router
@TheOneButlerian @EricNewcomer This is one of the things I am hearing about is people are trying to cobble these models together to get something that works, and everyone's experience seems to be different.
Probabilistic vs. deterministic.
@EricNewcomer Will never forget the day in negotiations class where we learned about maximizers vs. satisficers. TL;DR, even when maximizers have a higher marginal return, satisficers live a much happier life!
@EricNewcomer The second best one for everything
@EricNewcomer I've been wondering this for so long. Why would I choose Gemini, Copilot, ChatGPT, Claude, etc., beyond brand recognition?
@EricNewcomer Need a capability router! Building this in public @modelmeld
https://github.com/modelmeld/modelmeld
@EricNewcomer I think this will be a skill really needed in the future. Select the appropriate models and tools to minimize costs and maximize quality.
@EricNewcomer Stick to whatever the latest @OpenAI release is and you’re set. Not complicated at all.
No sense in chasing myths and fables.
@Jason @EricNewcomer @perplexity_ai @AravSrinivas Who even uses perplexity?
@EricNewcomer Eventually this gets abstracted away - right now it’s like using an iPhone that doesn’t know when to use WiFi, cell, the camera or mic and you have to tell it which one every time.
@EricNewcomer I feel like it’s just dependent on your needs and it’s subjective. It just depends on which products you prefer from the big ai companies.
@Jason @EricNewcomer @perplexity_ai @AravSrinivas This is exactly what i was looking for
@EricNewcomer that's too much thinking 4.8 opus for everything, get 3 max pro subs and never run out of credits 😅
@EricNewcomer Nobody asks which database their app is using. AI model selection should get to that same place eventually.
@EricNewcomer Genuinely the answer is probably the faster and cheaper one
@EricNewcomer the fatigue is real but the unlock is that for most tasks the model barely matters. pick one default, use it for everything, only switch on the long hard multi step stuff. the per task optimizing costs more time than it saves
@Jason @EricNewcomer @perplexity_ai @AravSrinivas @Jason bro, do you have a stake in the ASS company?
ASS = Aravind SrinivaS
Investor Jason suggested Perplexity's automatic routing as a fix
@EricNewcomer use @perplexity_ai and it will route automatically -- does a solid job at it. cc @AravSrinivas for background
as a consumer it's getting a little exhausting figuring out what model I need for every particular task
