Sabitlenmiş Tweet
Colin
2.4K posts
Colin
@0xtestpilot
Janitor @OasisProtocol | privacy/acc Prev: @ChainSafeth
Katılım Mayıs 2021
676 Takip Edilen1K Takipçiler
Two and a half years ago, we were three months from shutting down.
Today, this is live.
x.com/TACEO_IO/statu…
TACEO@TACEO_IO
The TACEO Network is live! A private execution layer for digital rails. The cryptographic infrastructure behind it already secures personal data for nearly 18 million people.
English
holy mother of slop companies
everything it recommended me was
-outdated
- super generic
- outright a bad idea
I hope my competition will use the AI cmo so I can crush them lmao
Okara@askOkara
Today we're introducing the world's first AI CMO. Enter your website and it deploys a team of agents to help you get traffic and users. Try it now at okara.ai/cmo
English
I noticed a sentiment change towards AI.
Before, engineers used to marvel in awe at this technology.
Nowadays, almost everyone acknowledges the speed, but utterly frustrated by the quality.
Fellas, I think humans evolved faster in the last 3 years than AGI.
English
Being in SF is like being in Wuhan right before the pandemic
Something is happening, it's gonna hit everywhere but so few people know it
English
Colin retweetledi
@hyperonline That esoteric memory management system you built? It will be completely obsoleted in 6 months by a native version that's way better 🎯
English
"You need to learn all these esoteric AI skills IMMEDIATELY or else blah blah permanent underclass"
This is completely retarded
Every "AI skill" people have learned over the last 3 years has been obsoleted 3 months later when a new model comes out
Remember 'ultrathink?'
Chain-of-thought prompting?
Goofy roleplay prompts? ("You are a genius software developer. Please oneshot this for me")
Heavy few-shot prompting?
Context engineering?
This is just going to keep happening
There are several hundred-billion-dollar companies whos job is to make this shit so easy to use that your mom could figure it out on her cellphone
That esoteric memory management system you built? It will be completely obsoleted in 6 months by a native version that's way better
The only "AI skills" that have been consistently useful this entire time are just "having good ideas" and "being able to communicate them clearly"
If you cannot do those things, then yah, you're going to be just as bad at AI as you were at everything else
_gabrielShapir0@lex_node
I'm increasingly sensing that "people who get really good at AI will survive & thrive" is likely a cope It doesn't require much skill, obviously it will also require less & less skill...I don't think any 'edge' you have today as a mere AI 'power user' will last long...
English
@ErikVoorhees Couldn't agree more. I'll never stop being surprised how little attn is dedicated to this topic
English
All the discussion about privacy in AI is such a win, few topics deserve more mindshare.
English
@itsolelehmann Its a great idea. Please post your results, though - cost seems like a limiting factor
English
social media is literally a slot machine and i just decided to break mine
im building an anti-dopamine webapp for my own X use.
here's my plan:
1. feed only opens 2x30 min per day
2. shows you interesting people to respond to, autoblocks ai slop comments (does research on every person commenting on your post)
3. hard lock before 10am and after 6pm
4. notifications come in one digest every 6 hours instead of constantly
5. before the feed opens you have to type what youre here to do. when youre done it tells you to leave
6. tracks creating vs consuming ratio
7. shows who you actually talk to vs who you just watch
the whole thing is based on one idea:
variable reward schedules are what make social media addictive.
I love to share my thoughts on social media and it's great to build connections and businesses
but man, it's also the worst attack on my daily feeling of happiness and balance
some people might be able to handle it, i can't
so I'm trying to build myself out of it :)
give me more ideas what to add
English
@VittoStack This looks super useful
x.com/0xtestpilot/st…
Colin@0xtestpilot
Memory is one of AIs biggest unsolved problems Inference is a commodity, but if you want to move around, you're often starting from zero Memory’s not portable. It’s not private. It’s barely even functional. ______________________________ There are technical reasons why this is a hard problem, providers also treat it as a moat. @OpenAI has a cross-chat memory feature & @AnthropicAI has search/archiving, but these are proprietary, closed-loop fixes They create something memory-like by continually trying to expand the context window In reality, it’s just building further dependency/lock-in Where does it leave us? → With a static, platform-dependent feature → When what’s needed is a portable, dynamic memory This is where crypto might provide an answer As @marco_derossi said, AIs should be local, replaceable, forgetful The vulnerability of the current setup is not just that data is stored centrally, but it’s processed/shared without trust. The solution is a user-owned layer that encrypts context while still making it usable by AI - decoupling memory from the model Projects working in this direction: -@PluralityWeb3 -@kinic_app -@ensue_ai -@memsync_ai -@rei_labs -@ekailabsxyz With better memory primitives, you can rent access to curated knowledge or monetize other expertise. The memory itself becomes the asset. Incredibly bullish on better memory solutions. Check out @markowifk's full piece on memory↓
English
We shouldn’t lose context every time we switch AI tools.
Introducing SharedContext ✨
- End-to-end encrypted
- Portable across devices
- Persistent across Claude, Cursor, Codex
- Shareable by link - same exact context
- Synced on the Blockchain
Only you can decrypt it. Fully Open Source and free.
"npm i -g ai-singlecontext"
Give it a try.
English
@LindyTasteful If I somehow attended this by mistake, would try to throw myself off the balcony
English
The LinkedIn-ification of American cities is kind of tragic
English