Zehra ^^ ✌ @bbozehra - Twitter Profile

about 2 hours ago

the witch had a knack for falling asleep on the couch. she had a bed, one all to herself, with a thick comforter and expensive sheets and pillows that could have swallowed her whole if she let them.

bbozehra's tweet photo. the witch had a knack for falling asleep on the couch. she had a bed, one all to herself, with a thick comforter and expensive sheets and pillows that could have swallowed her whole if she let them. https://t.co/v2qtXMjcAQ

0

10

Zehra ^^ ✌ @bbozehra

1 day ago

It’s going down to 60 tonight- perfect to sleep with the windows open and the fan on but instead we have to keep the windows sealed with the hvac on and hepa filters running. I don’t understand how the people on either side of her

0

1

0

20

bbozehra retweeted

beamnxw ./

@beamnxw

3 days ago

This paper is f*cking phenomenal A computer science paper unifies long-term and short-term memory management directly into the policy of AI agents The result: AgeMem exposes memory operations as tool actions, using 3-stage progressive RL and step-wise GRPO to optimize storage, retrieval, and context pruning The crazy part is how reinforcement learning teaches models when and what to remember Agents autonomously invoke ADD, UPDATE, SUMMARIZE, and FILTER tools during reasoning loops, cutting prompt token overhead while boosting accuracy across 5 benchmarks Most agent architectures rely on static heuristics or separate rule-based memory managers This system trains the core LLM policy to manage its own memory end-to-end Read the complete paper + article below Bookmark it for future reference

beamnxw's tweet photo. This paper is f*cking phenomenal

A computer science paper unifies long-term and short-term memory management directly into the policy of AI agents

The result: AgeMem exposes memory operations as tool actions, using 3-stage progressive RL and step-wise GRPO to optimize storage, retrieval, and context pruning

The crazy part is how reinforcement learning teaches models when and what to remember

Agents autonomously invoke ADD, UPDATE, SUMMARIZE, and FILTER tools during reasoning loops, cutting prompt token overhead while boosting accuracy across 5 benchmarks

Most agent architectures rely on static heuristics or separate rule-based memory managers

This system trains the core LLM policy to manage its own memory end-to-end

Read the complete paper + article below

Bookmark it for future reference

26

301

34

332

22K

bbozehra retweeted

RetroChainer

@RetroChainer

3 days ago

THIS IS WHERE "THE CLOUD" ACTUALLY LIVES - A ROOM OF DELL STORAGE ARRAYS YOU RENT A SLIVER OF EVERY MONTH. that clip is a real datacenter aisle, rack after rack of dell powerstore arrays, every hexagon panel packed with drives. this is "the cloud." not a metaphor, this room, plus the power, cooling and staff keeping it alive. when you pay for cloud storage, you rent a few gigabytes of these arrays. forever. the bill grows with every photo. here's the part they'd rather you not connect. these enterprise drives run on a schedule. data centers swap them out long before they fail, by the pallet. those decommissioned drives get tested and resold as recertified, the same 20tb enterprise disk for a fraction of new. so the exact hardware humming in this room ends up in a nas on your desk, minus the monthly rent. no cloud subscription, no data on someone else's array, no bill that grows with your library. the uncomfortable part isn't that the cloud is huge. it's that "the cloud" was always just this room, and you can buy a piece of it once instead of renting it forever. follow @RetroChainer for what the cloud doesn't want you to picture, and bookmark this before your next storage renewal.

23

145

36

65

33K

Zehra ^^ ✌ @bbozehra

2 days ago

@shmidtqq

0

8

Zehra ^^ ✌ @bbozehra

2 days ago

@Deepakk30081433 @grok does she have a youtube channel or just x

0

bbozehra retweeted

Skaly_Bull

@Skaly__Bull

2 days ago

The company that builds the AI writes none of its code by hand The engineer who built Claude Code said it on stage, not in a thread "Anthropic writes none of its code by hand at the company" While everyone sharpens a cleverer prompt he's already past the rung where prompts even matter Same weights you have open right now a clueless intern under a weak setup a senior engineer under a strong one The model was never the skill the system around it is Full breakdown below

15

82

25

51

10K

bbozehra retweeted

Psalter

@Psalteric

3 days ago

A truckload of empty aluminum cans is about 95% air. Nobody gets rich shipping air. That is why this machine exists. The lid seals the chamber, a hydraulic ram drives in from the side, and 15 seconds later a pile of cans leaves as a solid block of metal. Loose cans run roughly 30 to 50 kg per cubic meter. Baled, you get ten times that. Same truck, same diesel, ten times the metal. But here is the part nobody talks about. Baling is a control problem, not a crushing problem. Aluminum is elastic, so it pushes back. If the ram hits full pressure and releases at once, the bale swells and sheds cans for the rest of its life. The press has to hold pressure until thousands of thin walls buckle and lock into each other. The bale holds because the cans are tangled, not because the wire is tight. The bottleneck in recycling was never chemistry. It was geometry.

1

30

4

29

67K

bbozehra retweeted

Aman

@amandeep336

3 days ago

Just turned $105 into a quick W on Polymarket 🔥 BTC Up or Down 5m… hit that late-game pump and cashed. MetaMask → Polymarket → green candles only 📈 Who else is printing on these 5-minute markets? #Polymarket #Bitcoin #Crypto @Polymarket @PolymarketTrade @Gautamguptagg

1

7

1

0

543

Zehra ^^ ✌ @bbozehra

2 days ago

I know how overwhelming those feelings can be because I've experienced many of them myself. I'm not a therapist or a professional coach, just another teenager who has spent a lot of time learning about self improvement, discipline, productivity, and personal growth while trying