Pratham

17 | Building @trycatalystai

[Pratham @prathamarchives]

this is exactly what's gonna happen

[Jay @jayair]

Introducing Opus 4.9 — 10% smarter, 2x more expensive

Definitely NOT Fable

[Pratham @prathamarchives]

been seeing so many startups solving the "memory" "context" layer, and building knowledge bases, second brain and what not

and we are not far away that these AI giants would ship there large native memories and stretch context windows its already happening

models are going getting good enough at tasks, so good that government had to restrict the public access lol

imo after AI knows how to do the task the next layer is who it is doing for (identity layer), and how it should be done, the ability to judge the task if its "good" or "bad" according to user's preference (judgement layer)

@trycatalystai is giving your agent portable memory + identity + judgement

means now your ai agent knows and remember also knows who is acting on behalf of and has ability to evaluate every output against your standards

[Naval @naval]

The only truly wasted time is time spent wishing you were somewhere else.

[Pratham @prathamarchives]

Your AI sucks, models are being commoditized, your advantage is to leverage what you already have your taste, your creativity, your ability to see an output and say "this is shit"

the problem with AI is that it is trained on vast public data so it averages the internet and becomes sloppy, what if? your AI had your taste and actually knows what "good" means and don't stop there but as well as learns and improves over the time you use it

i just shipped an open source local-first project around the same thesis that let your agent extract your how you talk, what you actually like, what you reject, how you work through your existing AI sessions/workspaces and builds you a Creative Brain locally so your ai no longer produces a average "slop" and writes its own skills to keep improving and learning through your feedback so it compounds

just copy the repo link below paste it into your ai agent and say help me install this and it will guide you through

https://t.co/eao2FE9Y0P

[Pratham @prathamarchives]

RT @prathamarchives: Your AI sucks, models are being commoditized, your advantage is to leverage what you already have your taste, your cre…

[Pratham @prathamarchives]

@hipreeeet @ns peak

[Catalyst @trycatalystai]

Catalyst is now open-source. It lets an agent build a local Creative Brain from your AI sessions, workspaces, notes, drafts, corrections, and feedback.

https://t.co/gtFBKzcILj

[Pratham @prathamarchives]

@alexsllater do it right now

[Pratham @prathamarchives]

"this model is powerful enough that access to it may affect national security"

[Pratham @prathamarchives]

@robj3d3 now you don't

[Pratham @prathamarchives]

@AskVenice now you can't

[Pratham @prathamarchives]

@RexhaRexhaRexha 😂

[Pratham @prathamarchives]

@MEvsIRS not sure about AGI but AIG(any indian guy) is def coming for your job

[Pratham @prathamarchives]

@mil000 itt iss

[Pratham @prathamarchives]

@heyenzoexe i need thissss

[Pratham @prathamarchives]

@michellezfr @AruvalHQ sounds cool

[Pratham @prathamarchives]

access to intelligence may become unequal

[Pratham @prathamarchives]

this was not on my list

imo what's coming in future are 3 layers

you'll see public baseline models, good, cheap, fast for tasks like writing, coding, basic agentic tasks

and premium models will be metered, it will be expensive you'll see usage caps. dynamic pricing, priority tiers, maybe verified identity for access

and regulated/licensed frontier access, for some models like mythos it'll be enterprise only, US based access only, not being accessible to certain regions and like task restrictions

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[Pratham @prathamarchives]

@clear_graphics this look clean, template