permaXag

I enjoy macro and finance

[permaXag @permaXag]

@LoganARobison Hawaiian Bros in Houston? Pretty 🔥

[permaXag @permaXag]

@Paddy_Stash What’s your price target?

[permaXag @permaXag]

@robbertleusink Camden lives matter

[permaXag @permaXag]

@samschmitz @notadampaul Go to the beach in NJ 😅

[The Assembly @InTheAssembly]

While everyone gets ready for the Anthropic IPO, the real asymmetric upside in AI is sitting in 5 small-cap names almost nobody is covering.

Here they are:

[permaXag @permaXag]

@KingdomAmalland @mmjukic True

[permaXag @permaXag]

@yacineMTB Why does nobody talk about Manus?

[permaXag @permaXag]

@WillManidis The means already exist. It’s in the student

[permaXag @permaXag]

@fgonzalez1978 @FlaglerMuseum Yes

[Ricker @0xRicker]

https://t.co/Gnlkqlhll7

[permaXag @permaXag]

@thmoneycircle @GenZMultifamily Best neighborhood race map

[Dickie Bush 🚢 @dickiebush]

This week, I spent 8 hours reading 50+ articles about AI.

And I learned more in 1 day than most will in an entire year.

Highly recommend you read these 8:

[self.dll @seelffff]

10 repos that cut your ai agent token bill by up to 80%

1. microsoft/LLMLingua → cuts prompt size by up to 95%
compresses prompts before the api call. 20x compression.
published at EMNLP + ACL. near-zero quality loss.
6,100 stars
https://t.co/jUEM9PBZfj

2. mem0ai/mem0 → replaces full conversation history in context
stores what matters. retrieves only what's needed.
10,000 token history → 200 token memory. per agent.
54,800 stars
https://t.co/d0gT5owKfb

3. BerriAI/litellm → routes each call to the cheapest model
simple task → haiku. complex task → sonnet.
tracks cost per agent, per call, per day.
45,700 stars
https://t.co/ZigT7n1cbE

4. run-llama/llama_index → replaces sending full documents
rag: 100-page doc → 3 relevant chunks → same answer.
98% fewer tokens per query.
49,100 stars
https://t.co/D9sJa9aHeB

5. chroma-core/chroma → replaces keyword search in full context
vector store. finds the closest match. feeds only that.
50-200 tokens per query instead of thousands.
27,800 stars
https://t.co/pxMb3jXw0K

6. letta-ai/letta → replaces infinite context window crashes
paged memory for agents. loads only relevant memory.
stops your agent from hitting limits and retrying.
22,400 stars
https://t.co/FA0mtyHTMx

7. guidance-ai/guidance → cuts output token bloat by 30-50%
structured generation. constrains model output natively.
no more 100-token prompts to get json back.
21,400 stars
https://t.co/9VeJCmZrTk

8. Aider-AI/aider → replaces pasting entire codebases
builds a repo map. sends only files relevant to the task.
not your whole project. just what the agent needs.
44,300 stars
https://t.co/Oc185wtYIm

9. openai/tiktoken → count tokens before you send
know the exact cost before the api call happens.
not after the bill arrives.
18,100 stars
https://t.co/87KUhzS0gX

10. simonw/ttok → hard cap on what gets sent
cli tool: count tokens, truncate to budget limit.
pipe any text in. get truncated output back.
389 stars
https://t.co/hWDhCnbOG3

most agents are expensive not because the model is expensive.

because nobody checked what was being sent to it.

[Dwarkesh Patel @dwarkesh_sp]

Did a very different format with @reinerpope – a blackboard lecture where he walks through how frontier LLMs are trained and served.

It's shocking how much you can deduce about what the labs are doing from a handful of equations, public API prices, and some chalk.

It’s a bit technical, but I encourage you to hang in there - it’s really worth it.

There are less than a handful of people who understand the full stack of AI, from chip design to model architecture, as well as Reiner. It was a real delight to learn from him.

Recommend watching this one on YouTube so you can see the chalkboard.

0:00:00 – How batch size affects token cost and speed
0:31:59 – How MoE models are laid out across GPU racks
0:47:02 – How pipeline parallelism spreads model layers across racks
1:03:27 – Why Ilya said, “As we now know, pipelining is not wise.”
1:18:49 – Because of RL, models may be 100x over-trained beyond Chinchilla-optimal
1:32:52 – Deducing long context memory costs from API pricing
2:03:52 – Convergent evolution between neural nets and cryptography

[permaXag @permaXag]

@brim006 Screw the Ash Borer

[Rahul @sairahul1]

Karpathy didn't make a course.

He made THE course.

3 hours. Free.

Tokenization. Attention. Hallucinations. Tool use. RLHF. DeepSeek. AlphaGo.

Every behavior you've ever wondered about in an LLM - where it comes from, why it exists, how it was engineered.

The gap between engineers who understand this and engineers who don't isn't technical depth.

It's the ability to conceive of entirely different things.

[permaXag @permaXag]

@Tusslindo Cost of living isn’t coming down… ever 😂😂

[permaXag @permaXag]

@Say_JB_247 @ContrarianCurse If you someone was a single handicapper before and play infrequently now it’s definitely possible

[permaXag @permaXag]

@pronounced_kyle Yes. 15 “no go” 485 “go”

[permaXag @permaXag]

@hvsperus @DJBranham It’s not about flying to Philly. It’s getting to State College. Philly and Newark airports are both 3.5-4h drive away