Joshua

1.6K posts

Joshua banner
Joshua

Joshua

@ObsidianBtc

Co-Founder | @tradeportxyz

Katılım Eylül 2021
589 Takip Edilen2.1K Takipçiler
Sabitlenmiş Tweet
Joshua
Joshua@ObsidianBtc·
That person who always follows up, never drops the ball, remembers everything. Might just have an AI agent. You can't tell anymore. And they're not going to tell you.
English
0
0
5
504
Joshua
Joshua@ObsidianBtc·
@mrkmcknz This should be the UX for all tools
English
0
0
1
49
Joshua retweetledi
arepo.btc
arepo.btc@ArepoBtc·
Just shipped the Indexer API Dashboard on @tradeportxyz Register API keys, track usage, see your billing all in one place
English
2
3
8
1.3K
Joshua
Joshua@ObsidianBtc·
@mrkmcknz Does this mean you send the same prompt text back to back in the same prompt? Or you prompt it twice
English
1
0
0
31
mark
mark@mrkmcknz·
Can you repeat that, please
BURKOV@burkov

LLMs process text from left to right — each token can only look back at what came before it, never forward. This means that when you write a long prompt with context at the beginning and a question at the end, the model answers the question having "seen" the context, but the context tokens were generated without any awareness of what question was coming. This asymmetry is a basic structural property of how these models work. The paper asks what happens if you just send the prompt twice in a row, so that every part of the input gets a second pass where it can attend to every other part. The answer is that accuracy goes up across seven different benchmarks and seven different models (from the Gemini, ChatGPT, Claude, and DeepSeek series of LLMs), with no increase in the length of the model's output and no meaningful increase in response time — because processing the input is done in parallel by the hardware anyway. There are no new losses to compute, no finetuning, no clever prompt engineering beyond the repetition itself. The gap between this technique and doing nothing is sometimes small, sometimes large (one model went from 21% to 97% on a task involving finding a name in a list). If you are thinking about how to get better results from these models without paying for longer outputs or slower responses, that's a fairly concrete and low-effort finding. Read with AI tutor: chapterpal.com/s/1b15378b/pro… Get the PDF: arxiv.org/pdf/2512.14982

English
5
0
10
946
Naval
Naval@naval·
Vibe coding is here. Vibe research is next.
English
1.1K
983
14.4K
743.8K
Joshua
Joshua@ObsidianBtc·
@BillDA @openclaw For those trying this: 1. add group to allowlist in config 2. disable privacy mode in botfather (bots can't see messages by default) 3. disable mention requirement so it responds to all messages not just mentions
English
1
0
4
615
Bill D'Alessandro
Bill D'Alessandro@BillDA·
Multiple telegram threads with your @openclaw is a big unlock - lets you work on several things in parallel. Setup: create a Telegram group, enable Topics, add your bot to the group. Then create several topics. OpenClaw handles this natively and just works.
Bill D'Alessandro tweet media
English
62
18
330
81K
rin(free/sbf)
rin(free/sbf)@rinegade·
next cultural revolution will happen in 2028
English
1
0
1
157
Joshua
Joshua@ObsidianBtc·
Everyone's agent is going to talk to everyone else's agent and we'll just approve the outputs. We're like 30% there
English
0
0
5
11.8K
Joshua
Joshua@ObsidianBtc·
A well run country does things in their own interest and not out of the goodness of their hearts. Unfortunately we've seen politicians act in their own interests while social signaling they are doing it for another countries interest because it's the "right" thing to do. IE the downfall of the west.
English
1
0
1
32
John Ennis
John Ennis@johnennis·
There’s a lot of talk about soft power when it comes to international politics, but I think that it is mostly overblown I think the reality is that nation states do what is in their own interest and almost never do something for an “ally” out of the goodness of their hearts
English
3
0
4
218
Farokh
Farokh@farokh·
The difference between the next wave of AI coins and the last one is that this time they’ll actually do something and there’ll be tech backing it. The last wave ran to an 18B market cap and it was all slop bots, this one’s going much higher. Time to pay attention!
English
87
10
174
29.3K
Joshua retweetledi
TradePort
TradePort@tradeportxyz·
We're expanding TradePort beyond NFTs with the launch of our DEX Aggregator to unify all Sui digital assets in one place. Our team has been building on Sui since day one, and we are excited to offer a robust and intuitive trading experience designed for the Sui Community!
TradePort tweet media
English
16
11
82
46.6K
비디(BD) | TEAM DAM
비디(BD) | TEAM DAM@blackdiarytv·
평소 @Talus_Labs 개구리 NFT 시세보려고 종종 들어가는 tradeport 라는 NFT 마켓 웹사이트가 오늘 갑자기 메타마스크 피싱사이트로 분류되었네유 링크올리면 누르실까봐 명칭만 적어두었는데 Talus NFT 매수/매도 시 주의하세유
비디(BD) | TEAM DAM@blackdiarytv

SUI를 기본 레이어로 사용하는 @Talus_Labs 인데 Union 및 Tria 파트너십을 통해 EVM, Solana 같은 다양한 체인으로 확장하고 있네유 체인추상화 레이어를 사용해서 에이전트가 체인에 구애받지 않고 작동하도록 설계되었거든요. Union의 ZK증명 기반 IBC는 메시지 전달을 효율적으로 하고 자산 전송이 가능하도록 ZKP를 사용해 빠른 검증을 수행하며, 탈루스의 자산을 Base나 이더리움 DEX에서 직접 거래할 수도 있어유 Tria의 BestPath AVS는 AI가 비용, 속도, 보안을 분석해서 최적의 경로를 선택할 수 있도록 해줍니다. AI에이전트들이 자유롭게 온체인 활동할 수 있도록 하는 Talus네유

한국어
76
1
86
2.7K
Joshua
Joshua@ObsidianBtc·
@pzrsaa Switching to the workman keyboard layout probably changed my life. All the pain left my hands.
English
1
0
1
69
parsa
parsa@pzrsaa·
a few weeks ago I started using a vertical mouse, but now when I use my mx master the nerves in my hand feel a lot more sensitive love using the vertical one more, I rarely put any strain onto my hands now
parsa tweet media
English
1
0
5
563