Large Language Model (LLM) Social Insights and Sentiment Analysis

my favourite and laziest way to use ai is to simply copy and past entire conversations into the llm and tell it to action it. stop overthinking ai

5/13/2026 16:18

5.9K

6

1

115

24

here we are! the first known zero-day developed with ai just got caught by google. someone used an llm to spot a flaw in an open-source admin tool. then had the model write a python script that bypasses 2fa on it. then started using it on real targe...

5/13/2026 12:00

837

0

1

2

the impact on markets of how fast information spreads now on social media combined with consensus through LLM usage really cannot be overstated the hot ball of money is just massive

5/13/2026 03:23

33K

120

29

549

47

guys dont spend all your llm tokens on day 1 of hot emin

5/12/2026 19:35

1.3K

15

10

83

0

Wasting LLM credits by sending a "thank you!" is an extremely underrated AI flex

5/12/2026 16:51

818

5

0

23

0

In 2026, not making your startup pitch deck / materials LLM-digestible is insanity. Materially increases the chance of insta-pass. LLM-first everything. I'm going to be so much better prepared if I can have a conversation with my LLM about what you...

5/12/2026 09:41

30K

55

6

218

112

Have you ever thought about what most people actually want from LLM tokens? Do they want the tokens? or do they want the outcome? 🤔

5/12/2026 00:10

641

4

0

7

0

🤔 who else is hitting reload constantly on 5 different llm providers' cost/usage page trying to understand which model is costing what in realtime? 🙋‍♂️

5/11/2026 18:54

2.5K

9

1

40

1

LLM Wikis + HTML Artifacts is one of the most underrated combos in AI workflows •⁠ ⁠Wiki captures everything agents need to know •⁠ ⁠Artifact sits on top & makes it interactive Zero inbox. Live research. Dynamic interfaces. No website needed....

5/11/2026 15:15

49K

11

4

86

43

The market for private AI is only a fraction of what it will become. Venice has built the best app (by far imo) for private multi model LLM usage, driving continued exponential acceleration in Venice app usage, $VVV burns, and $DIEM inference usage. ...

5/10/2026 23:42

3.9K

5

1

67

11

Alibaba just dropped a multimodal LLM that understands documents without OCR. No text extraction pipeline. No preprocessing. Just point it at a PDF and it reads it like a human. #5 on GitHub trending today. https://t.co/WPD7wJYmJK

5/10/2026 20:15

47K

7

4

67

41

Every AI agent security claim you've seen is based on static benchmarks. Known attacks. Fixed inputs. Check the boxes. Ari Medvinsky from @getfailsafe says that's useless, real adversaries adapt. So he built Attack Bench. An LLM-powered red tea...

5/9/2026 12:50

865

0

3

0

Schizoanalysis arrived decades too early, but it may be exactly the psychology we need for LLMs If there is such a thing as “LLM psychology,” it will map less onto Neo-Freudian interiority than onto Guattari’s machinic assemblages, semiotic flows, d...

5/9/2026 08:01

2.4K

3

2

57

19

This is pure gold. Marc Andreessen's custom system prompt that makes any LLM 10x smarter. You'll want to save this: https://t.co/vCqNGnX3e9

5/7/2026 19:30

494K

185

641

5.8K

14K

1.5b LLM tokens spent in 9 days 5b more tokens and then up only

5/5/2026 18:39

809

2

1

18

0

Jameson Pickett, Head Quant at @TheoriqAI, explains how AI can prevent the next exploit: "On Resolv, curator automation poured user capital INTO the compromise before a human came in hours later. But @yieldsandmore on Twitter knew in 5 minutes. "A...

5/5/2026 17:54

888

2

12

0

New Silicon Data Index just dropped --> The "LLM Token Expenditure" Index is a daily benchmark designed to measure the effective expenditure level of the actively traded broad LLM market, measuring price per million tokens. https://t.co/Rq5kxlIKqp

5/5/2026 17:53

9.3K

7

2

64

15

How I write these days: - By hand if it's short (like on twitter). Actually spend time to make it short. - If it's long, I have a skill distilled from all my writings in the past; I then give it bullet points of what I really want to say and have t...

5/5/2026 17:25

2K

1

0

12

4

The best LLM setup there is right now: Frontend: Opus 4.7 Backend: GPT-5.5 Debugging: Sonnet 4.6 Research: Grok 4.2 Writing: GPT-5 Mini Image: Gemini 3.1 This is all you need.

5/5/2026 16:15

96K

66

42

1.1K

652

The EU 🇪🇺demands access to Anthropic's "Mythos" LLM, which is not publicly available. They are concerned their banks run on crappy legacy software full of exploitable bugs (true), and only Mythos can save them (false). If the EU cannot access good...

5/5/2026 08:19

2K

6

4

40

3

ANTHROPIC PAYS $750K FOR LLM ENGINEERS AND STANFORD TAUGHT IT FOR FREE IN 1 HOUR https://t.co/RTFaIe41C2

5/4/2026 06:45

50K

16

15

174

156

The Ai heads and vibecoders can spend the next 10,000 years on their little LLM bots and they’ll never create anything anywhere remotely near the value of Bitcoin. In fact for a tech that has half of the worlds 20-30 year olds sinking billions of ho...

5/2/2026 14:45

13K

17

8

177

9

RAG retrieves. LLM Wiki builds and remembers. Instead of searching your docs fresh every query, it compiles everything into a structured, interconnected wiki that updates incrementally as you add more. > Four-signal knowledge graph connecting do...

5/2/2026 13:00

47K

13

2

72

42

me and dario were discussing how decentralized AI could be the future of AI, compute and LLM token economies https://t.co/h97OEBm1DJ

4/30/2026 18:39

577

3

0

12

0

How to think about LLMs & harnesses LLM - The CPU / hardware Harness - The OS

4/29/2026 23:25

568

4

0

4

0

📊 when the @clawdbotatg agents are putting in the work but staying as efficient as possible (most of the time) 🏦 huge shoutout to our @bankrbot LLM provider and pretty graphs https://t.co/0aRAp6og2u

4/29/2026 22:58

3.3K

4

14

61

1

I have spent $450 in llm tokens in 2 days It's a great app for me tho https://t.co/kyaWzHsdFi

4/29/2026 16:21

690

3

0

30

0

the bait is incredible today, but I will stay in my lane, and spend LLM tokens instead https://t.co/EtpRYqRDBI

4/29/2026 15:30

367

3

0

25

0

thinking of building a service that lets you prove your LLM API calls e.g. "this piece of text was generated by opus 4.7 in response to this prompt" would you want to use this?

4/29/2026 04:14

5.2K

13

0

38

6

I see all this cool AI agent thinking, enrichment, etc and I think cool but how is anyone affording this? I mean I must have spent 100 hours racking my brain on if I should spend an LLM call for agent to decide to respond in a group chat and when to...

4/28/2026 20:27

6.2K

29

2

79

6

lately been obsessed with reading LLM generated books in flights (without wifi) you get to nerd & mix-match any topic you like eg. i generated a book on all financial instruments over last 500 years taking ref from 3-5 books + personalised &...

4/27/2026 15:14

3.6K

7

2

64

33

There’s so much inefficiency in this market. The proliferation of AI for research actually means less people think outside the box or dig beyond what an LLM guides them into. Been looking at something recently and all 3 top model providers barely sc...

4/27/2026 14:13

5.3K

8

0

68

13

Really interesting site essentially aggregated >85M scientific papers and is queryable via an LLM interface. Can think of it as AI PubMed. https://t.co/KMGOXOvet1

4/27/2026 13:09

7.3K

2

4

68

84

Deepseek V4 feels like a golden retriever if it was an LLM model

4/26/2026 00:12

2.5K

5

0

21

1

for adhd people LLM-based agents are categorically better to communicate with because you don't need to worry about talking linearly as much as just getting the right keywords into the prompt

4/25/2026 23:55

691

0

9

2

A SINGLE CLAUDEmd FILE JUST TURNED ANDREJ KARPATHY’S CRITIQUE OF LLM CODING INTO A PRACTICAL FIX FOR CLAUDE CODE. Four simple rules, one install, and a cleaner way to stop overcomplication, silent assumptions, and useless rewrites. https://t.co/rX0e...

4/24/2026 21:15

44K

7

5

91

77

So what actually happens to jobs and LLM shops when all the semi raws are truly capped out for 6months in 2H26?

4/24/2026 14:09

2.7K

1

0

7

3

just ask ai man, claude can do it, talk to an llm, did you try asking gemini, chatgpt can generate it, use claude code, did you ask grok, glean will def know it, just use the cowork dude

4/24/2026 00:03

82

0

4

0

the lion does not concern himself with the opinion of an LLM. the lion hit his daily session limit. the lion waiting for the reset

4/23/2026 23:52

52

0

6

0

Some definitions Agent: An LLM that calls tools in a loop Harness: Software that constrains the LLM, manages memory, packages skills and tools Products: Multiple Agents packaged in a harness

4/23/2026 13:17

2.1K

1

12

5

LLM Psychosis is the first named intersection of abnormal psych and AI. It’s delusional subtypical schizophrenia, without violence. The second is Claude Code Mania, which is Hypomanic Episode with Computer.

4/23/2026 02:03

9.1K

12

7

173

15

Karpathy's LLM-Wiki and Garry Tan's Gbrain concept just became a real product. Membase - memory, wiki, auto-update, and live sync with agents. Your second brain, a few clicks away. https://t.co/qbTQqhoFef

4/22/2026 13:45

54K

12

4

84

42

After blasting into the semi sector to go long the full 9 yard of AI stuff (CPU, GPU, ASIC, XRAM, LLM, etfs) with the thesis that we are chronically short the raws to power AI + AI demand continues to rock, where would you put incremental dollars to ...

4/22/2026 07:34

3.6K

5

0

16

3

LLM PSYCHOSIS https://t.co/O3b7czhNwh

4/22/2026 03:57

25K

72

105

853

21

optimising for language based interest clusters, filtered by large language model, when a major slop-attack vector is [compounded by] language models, is a symptom of misunderstanding the multifactored systems at play

4/22/2026 00:27

2.6K

1

0

23

3

Working with a LLM agent, and it said 'generate 8 keyframes with a new model, animate, stitch, add audio. Let me do it all in one job since I know the process now.' "Since I know the process now"

4/21/2026 13:31

168

1

0

1

0

If you've used an LLM for a while and written personal things, try this prompt: "Given everything you know about me, tell me a historical fact I've probably never heard about that would absolutely blow my mind."

4/21/2026 02:58

5.1K

7

1

17

6

talking to your llm all day is just elitist doomscrolling

4/21/2026 00:17

6.4K

12

6

150

4

I've navigated the crypto social space my whole life and I'm doing a @karpathy-style LLM wiki to predict the next idea maze move to make Referenced: @context, @farcaster_xyz, @foundation, @tryShowtime, @zora , @giggles_app, @frenzydotfun, @believeap...

4/20/2026 23:46

2.2K

12

0

21

1

Membase turned Karpathy's LLM-Wiki and Garry Tan's Gbrain concept into a real working product. Memory, Wiki, auto-update, and live sync with agents. Your second brain, a few clicks away. https://t.co/qbTQqho7oH

4/20/2026 20:45

2.8K

15

4

39

29

I don’t know what LLM’s you people are using, but for a few days now mine are nonstop slop and complete fabrications. Completely unusable.

4/20/2026 12:28

4K

18

0

48

1

it’s 2016 tove lo just came on at the house party you’re at you have no idea what an LLM is life is good

4/17/2026 22:13

1.5K

2

1

29

0

An LLM operating a GUI as fast as a person. Not a demo. Not sped up. Just surreal. https://t.co/bXX6NkIDwV

4/17/2026 16:45

48K

14

0

64

6

Your Favorite LLM™ rotates the wheel one single revolution more but it keeps rolling in the same direction. the flintknapper’s wife wasn't happy when the Bronze Age hit (he had to go out and finance a forge with zero percent interest) but he paid t...

4/16/2026 19:05

940

0

1

0

STOP PAYING FOR LOVABLE, BOLT, AND REPLIT. GITLAWB PLAYGROUND IS 100% OPEN SOURCE, POWERED BY OPENCLAUDE AND BANKR AS THE CRYPTO-NATIVE LLM GATEWAY. SAME THING BUT FREE. https://t.co/2eTFEybW5S

4/16/2026 16:45

52K

30

38

208

100

THIS GUY BUILT A MULTIPLAYER, COMPLETELY LLM-DRIVEN GAME https://t.co/Oyu02qBIfk

4/15/2026 16:15

48K

18

1

103

9

WECLONE TURNS CHAT HISTORY INTO YOUR AI TWIN - Fine-tunes LLM on your Telegram history (style, tone, vocab + images) and strips personal data - Deploys as live chatbot on Telegram/Discord/Slack, fully local - 100% open source (16.4K stars) on GitHub...

4/15/2026 15:15

46K

7

3

55

21

THIS IS A FULLY LLM DRIVEN MULTIPLAYER GAME WHERE YOU MANAGE AI AGENTS INSTEAD OF CLICKING https://t.co/Oyu02qBIfk

4/15/2026 13:15

45K

12

4

68

13

The future of business will be who can communicate with their llm better

4/14/2026 15:58

5.7K

7

2

57

0

Commercial LLM’s close to unusable roght now. The “golden era” for retail seems to be over already. They hooked you proper, now you gotta pay up.

4/14/2026 12:18

3.8K

11

1

49

1

llm psychosis is the new adhd

4/14/2026 02:15

5.9K

23

8

102

4

Sector:	AI Agents
Blockchain:	Solana

Free

Premium

Large Language Model (LLM)

$0.0₃ 165

Sentiment

Description

#5229

Social Statistics

Featured Tweet

Resources

About

Legal

Socials