# ![@HendrixLiu29208 Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::1752956949828890624.png) @HendrixLiu29208 Hendrix Liu

Hendrix Liu posts on X about ai, llm, yc, sf the most. They currently have [---] followers and [--] posts still getting attention that total [------] engagements in the last [--] hours.

### Engagements: [------] [#](/creator/twitter::1752956949828890624/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1752956949828890624/c:line/m:interactions.svg)

- [--] Week [-------] +3,143%
- [--] Month [-------] +240,804%
- [--] Months [-------] +6,397%
- [--] Year [-------] +8,630%

### Mentions: [--] [#](/creator/twitter::1752956949828890624/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1752956949828890624/c:line/m:posts_active.svg)

- [--] Week [--] +320%
- [--] Month [--] +1,300%
- [--] Months [--] +289%
- [--] Year [--] +132%

### Followers: [---] [#](/creator/twitter::1752956949828890624/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1752956949828890624/c:line/m:followers.svg)

- [--] Months [---] +44%
- [--] Year [---] +146%

### CreatorRank: [---------] [#](/creator/twitter::1752956949828890624/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1752956949828890624/c:line/m:influencer_rank.svg)

### Social Influence

**Social category influence**
[technology brands](/list/technology-brands)  [stocks](/list/stocks)  [finance](/list/finance)  [social networks](/list/social-networks)  [currencies](/list/currencies)  [cryptocurrencies](/list/cryptocurrencies) 

**Social topic influence**
[ai](/topic/ai), [llm](/topic/llm), [yc](/topic/yc), [sf](/topic/sf), [prompt](/topic/prompt), [for the](/topic/for-the), [$500k](/topic/$500k), [500k](/topic/500k), [developers](/topic/developers), [$googl](/topic/$googl)

**Top accounts mentioned or mentioned by**
[@keywordsai](/creator/undefined) [@openai](/creator/undefined) [@traeai](/creator/undefined) [@owengretzinger](/creator/undefined) [@anthropicai](/creator/undefined) [@awscloud](/creator/undefined) [@andydy42](/creator/undefined) [@dominikkoch](/creator/undefined) [@langfuse](/creator/undefined) [@braintrust](/creator/undefined) [@usenotra](/creator/undefined) [@hyperboundai](/creator/undefined) [@jacobilin](/creator/undefined) [@mistralai](/creator/undefined) [@shack15sf](/creator/undefined) [@glitchphoton](/creator/undefined) [@hendrix56915claude35sonnetvsgpt4owhichisbetterf4c4fe3a8f16](/creator/undefined) [@googlecloud](/creator/undefined) [@azure](/creator/undefined) [@alexdanilo99](/creator/undefined)

**Top assets mentioned**
[Alphabet Inc Class A (GOOGL)](/topic/$googl) [Braintrust (BTRST)](/topic/braintrust) [Uber Technologies, Inc. (UBER)](/topic/$uber) [DoorDash Inc. (DASH)](/topic/doordash)
### Top Social Posts
Top posts by engagements in the last [--] hours

"@jacobilin Hey Jacob Congrats on your Product Hunt launch My team and I just upvoted Glowby Genius. Would love to learn more about it - lmk if you're up for a quick chat with our team"  
[X Link](https://x.com/HendrixLiu29208/status/1770579765663322410)  2024-03-20T22:34Z [--] followers, [---] engagements


"New @MistralAI model Mistral-7B-v0.2 - 32K context window (instead of 8k) - Theta (RoPE) = 1e6 - No sliding window Try this new model on / @keywordsai Mistral just announced at @SHACK15sf that they will release a new model today: Mistral 7B v0.2 Base Model - 32k instead of 8k context window - Rope Theta = 1e6 - No sliding window https://t.co/iAuEUEOw5K Mistral just announced at @SHACK15sf that they will release a new model today: Mistral 7B v0.2 Base Model - 32k instead of 8k context window - Rope Theta = 1e6 - No sliding window https://t.co/iAuEUEOw5K"  
[X Link](https://x.com/HendrixLiu29208/status/1771685624287772686)  2024-03-23T23:49Z [--] followers, [--] engagements


"@GlitchPhoton Got it. Maybe you could try @keywordsai to help you do that. It's free now You can also get $20 credits if you book a demo with us :)"  
[X Link](https://x.com/HendrixLiu29208/status/1771977621829726222)  2024-03-24T19:09Z [--] followers, [--] engagements


"I wrote a blog yesterday comparing @AnthropicAI 's claude-3.5-sonnet and @OpenAI 's gpt-4o. TL;DR Claude-3.5-Sonnet excels in context precision faithfulness and readability. GPT-4o is faster and better for voice AI startups due to lower latency and quicker time to first token (TTFT). Check it out here: https://medium.com/@hendrix_56915/claude-3-5-sonnet-vs-gpt-4o-which-is-better-f4c4fe3a8f16 https://medium.com/@hendrix_56915/claude-3-5-sonnet-vs-gpt-4o-which-is-better-f4c4fe3a8f16"  
[X Link](https://x.com/HendrixLiu29208/status/1804255420611203318)  2024-06-21T20:49Z [---] followers, [---] engagements


"Top [--] deals for early-stage startup founders: 👇 Cloud services: - @googlecloud: Up to $350000 in Google Cloud credits - @awscloud: Up to $100000 USD in AWS credits - @Azure: Up to $150000 (USD) in Azure credits https://foundershub.startups.microsoft.com/signup https://aws.amazon.com/startups/creditslang=en-US#packages https://cloud.google.com/startuphl=en https://aws.amazon.com/startups/creditslang=en-US#packages https://cloud.google.com/startuphl=en https://foundershub.startups.microsoft.com/signup https://aws.amazon.com/startups/creditslang=en-US#packages"  
[X Link](https://x.com/HendrixLiu29208/status/1819891531463192745)  2024-08-04T00:22Z [---] followers, [--] engagements


"LLM: - @OpenAI: $2500 USD in credits (via Azure) https://foundershub.startups.microsoft.com/signup https://foundershub.startups.microsoft.com/signup"  
[X Link](https://x.com/HendrixLiu29208/status/1819891534919061762)  2024-08-04T00:22Z [---] followers, [--] engagements


"@alexdanilo99 @magicpatterns Happy birthday🎂"  
[X Link](https://x.com/HendrixLiu29208/status/1826034942599639424)  2024-08-20T23:13Z [---] followers, [--] engagements


"Since the release of Llama [--] and Llama [---] we've seen a 3X surge in usage of @GroqInc models on @keywordsai. LLM developers are flocking to Groq's open-source models due to the incredible speed powered by a revolutionary chip - the LPU. 🤔 What is LPU LPU (Language Processing Unit) is a specialized chip designed to accelerate AI language tasks. Here's the tech behind it: [--] Sequential processing: Optimized for language tasks that require step-by-step processing unlike GPUs' parallel approach. [--] High compute density: Packs more processing power into a smaller footprint for faster LLM execution."  
[X Link](https://x.com/HendrixLiu29208/status/1826167144540983709)  2024-08-21T07:59Z [---] followers, [---] engagements


"It was great to hear from @kevinweil CPO of @OpenAI today. His speech helped me understand if an AI startup or idea is valid. "If you fear OpenAI's new model will crash your startup reconsider your idea. If you're excited about the new model you're likely on the right track." This resonates with @keywordsai. We're excited about every new model release. More powerful LLMs will lead to more AI products. Developers can't make great AI products without observability. We're confident we're on the right track - building the leading LLM monitoring platform for AI startups"  
[X Link](https://x.com/HendrixLiu29208/status/1826907095536701569)  2024-08-23T08:59Z [---] followers, [---] engagements


"Every day @Andydy42 and I engage in a friendly competition in our Discord channel. Our challenge To be the first to welcome new members and respond to their questions. We believe that providing swift support enhances our users' experience on our platform. By turning this into a game we've found an exciting way to motivate ourselves to offer the fastest possible assistance to our community members. Curious to see who wins our next round Join our Discord community and witness the action firsthand: https://discord.com/invite/KEanfAafQQ https://discord.com/invite/KEanfAafQQ"  
[X Link](https://x.com/HendrixLiu29208/status/1828882778332770573)  2024-08-28T19:50Z [---] followers, [---] engagements


"🚀New free SEO tool for bloggers & creators🚀 Rank higher on Google with AI-powered SEO-friendly descriptions. ✅ Free GPT-4o mini ✅ GPT-4o & Claude [---] Sonnet support ✅ Self-hosting option Try it now:"  
[X Link](https://x.com/HendrixLiu29208/status/1834041832713715970)  2024-09-12T01:30Z [---] followers, [--] engagements


"September was a remarkable month for @keywordsai Highlights: - 2x website traffic compared to August - Showcased our platform at GitHub HQ - Demoed at Product Hunt and got 200+ new sign-ups. - Launched Lab feature for testing prompts with dynamic test cases - Introduced Testsets for editing test cases with a spreadsheet-like interface - Major upgrade to the Prompts page better version control and team collaboration - Implemented LLM cost management to prevent unexpected expenses - Enabled users to customize LLM retries to avoid failovers - And much more (See details in comments)"  
[X Link](https://x.com/HendrixLiu29208/status/1841635022497054777)  2024-10-03T00:22Z [---] followers, [---] engagements


"Had an ankle ligament tear last month. But I'm now back on the court. ready to smash 🏸💪"  
[X Link](https://x.com/HendrixLiu29208/status/1849625339494007116)  2024-10-25T01:33Z [---] followers, [---] engagements


"I always heard people say Claude [---] Haiku is faster than Claude [---] Sonnet But is this really true After extensive testing in the @keywordsai model playground I've found that Sonnet sometimes outperforms Haiku in both speed and generation time.Contrary to popular belief Haiku doesn't have an obvious advantage in speed overall. If budget isn't your main concern I'd recommend choosing Claude [---] Sonnet for better performance"  
[X Link](https://x.com/HendrixLiu29208/status/1861620556736729423)  2024-11-27T03:58Z [---] followers, [---] engagements


"4. LangGraph: A LangChain extension for creating stateful multi-agent systems with planning capabilities"  
[X Link](https://x.com/HendrixLiu29208/status/1862130895035261085)  2024-11-28T13:46Z [---] followers, [--] engagements


"5. Swarm: OpenAI's lightweight framework for multi-agent orchestration"  
[X Link](https://x.com/HendrixLiu29208/status/1862130896142475358)  2024-11-28T13:46Z [---] followers, [--] engagements


"My savings update for 2024: $11k - $5k At @keywordsai we founders live on a minimum salary - just enough for rent food and basic necessities. We're keeping our salaries low because every dollar saved goes toward building our company's future. Why has my savings decreased this year - Unpaid for [--] months due to visa issues. - Unexpected ligament tear injury on my right ankle"  
[X Link](https://x.com/HendrixLiu29208/status/1863393299488538786)  2024-12-02T01:22Z [---] followers, [---] engagements


"2025 is the year of AI agents. But what is the foundation of an AI agent It's models Choosing the right model is crucial. A bad choice can ruin everything youve built. I just wrote a blog listing the best LLMs for different tasks in [----]. Use it as a guide to select the perfect model for your needs. Link👇"  
[X Link](https://x.com/HendrixLiu29208/status/1878236374316503301)  2025-01-12T00:23Z [---] followers, [----] engagements


"Open-source LLMs are a better choice for building AI agents. Here are [--] platforms that host open-source LLMs for developers: Which one should you choose for your agents [--]. Start with @FireworksAI_HQ and @togethercompute. They offer stable access to popular open-source models. [--]. Consider @OpenRouterAI for a unified API to access most market LLMs. [--]. Use @replicate to access developer-published models. More details 👇"  
[X Link](https://x.com/HendrixLiu29208/status/1882901569135620115)  2025-01-24T21:21Z [---] followers, [---] engagements


"Last week a customer switched from a competitor to us because he needed a prompt playground that supports tool calls. In [----] AI products will no longer be just simple prompts. They will be complex workflows built from external tools guardrails and multiple models. A basic prompt playground won't be enough. AI developers need an interactive way to test their workflows. They need a place to easily combine function calls prompts and models from different providers. @keywordsai playground provides exactly that. Here's a quick demo of our prompt playground:"  
[X Link](https://x.com/HendrixLiu29208/status/1901738253033804197)  2025-03-17T20:51Z [---] followers, [---] engagements


"11x has been accused of claiming customers it doesn't actually have. Even worse its product doesn't seem to be working. Many users have said the AI SDR frequently hallucinates in customer emails. Maybe it's time for AI engineers at 11x to consider @keywordsai for their internal LLM evals😁"  
[X Link](https://x.com/HendrixLiu29208/status/1904594227172192285)  2025-03-25T18:00Z [---] followers, [---] engagements


"used to cook every meal to save money. shopping prepping cooking cleaning it easily took [--] hrs a week. now we do meal delivery for the team. costs more but gives us time back to ship faster. early-stage tradeoff: save money or save time pick what helps you move"  
[X Link](https://x.com/HendrixLiu29208/status/1927133628217065509)  2025-05-26T22:43Z [---] followers, [---] engagements


"LLM Evals is now part of every conversation in Silicon Valley. At @keywordsai we process over 500M+ LLM logs monthly. We see that many developers struggle with reducing hallucinations making agents follow instructions and retrieving memory quickly and correctly. Want tips on solving these challenges We're hosting an event on RAG and LLM Evals with Ragie. Join us to discuss best practices and learn from top AI developers and startups in Silicon Valley. 🗓 Thursday June [--] 📍 Olfactory Brewing & Blending San Francisco Free beer and snacks included Check the link👇"  
[X Link](https://x.com/HendrixLiu29208/status/1937208445834518942)  2025-06-23T17:57Z [---] followers, [---] engagements


"so all the founding members of windsurf ended up with nothing🥲 kind of wild our whole team actually stopped using it a month ago. the product just fell too far behind other coding tools. BREAKING: OpenAI's acquisition of Windsurf is off. CEO and other members of the team are headed to Google DeepMind. Source: The Verge https://t.co/xyxHsq8Qs2 BREAKING: OpenAI's acquisition of Windsurf is off. CEO and other members of the team are headed to Google DeepMind. Source: The Verge https://t.co/xyxHsq8Qs2"  
[X Link](https://x.com/HendrixLiu29208/status/1943840508100718646)  2025-07-12T01:11Z [---] followers, [---] engagements


"theres zero stickiness in LLMs. people just pick whatevers best - no brand loyalty no switching cost. so LLM companies have to burn $$$ nonstop just to stay in the game. and yeah thats why theyre all racing to build apps on top. they kinda have to"  
[X Link](https://x.com/HendrixLiu29208/status/1947804853918380518)  2025-07-22T23:44Z [---] followers, [---] engagements


"Sam Altman's right. The next generation will grow up in a world where they are never the smartest "being" in the room instead of trying to compete with AI on logic and data we should focus on leveraging it to amplify our own human creativity and wisdom"  
[X Link](https://x.com/HendrixLiu29208/status/1948883522456068189)  2025-07-25T23:10Z [---] followers, [---] engagements


"me and @francchen have been using @Trae_ai for [--] months now and its been a pretty solid experience at first Trae wasnt as smooth as other AI IDEs like Cursor windsurf or Claude Code definitely felt early and a bit buggy. I ended up bouncing between Cursor and Trae since I like trying out new stuff lol but as Trae improved it started catching up to Cursor and Claude Code. still nothing really stood out until they launched SOLO SOLO honestly changed everything for us. as our codebase got bigger and more complex we needed a tool that could actually understand our codebase and pull in outside"  
[X Link](https://x.com/HendrixLiu29208/status/1952513691955724619)  2025-08-04T23:35Z [---] followers, 11.1K engagements


"@jihyuk_gentoo @OpenAI Congrats"  
[X Link](https://x.com/HendrixLiu29208/status/1955752143371296901)  2025-08-13T22:03Z [---] followers, [--] engagements


"Excited to be speaking at the @Trae_ai SF Meetup 🎤 Ill share how we built @keywordsai - from early experiments to supporting hundreds of AI teams today. 📍 San Francisco 📝 Sign up: https://luma.com/w5hwgdwx https://luma.com/w5hwgdwx"  
[X Link](https://x.com/HendrixLiu29208/status/1969126042334879805)  2025-09-19T19:46Z [---] followers, [---] engagements


"Past [--] hours in tech 👇 - @OpenAI dropped Sora [--] (next-gen video) - Partnered w/ @stripe for instant checkout in ChatGPT - @AnthropicAI launched Claude Sonnet [---] (best coding model yet) - @DoorDash rolled out a delivery robot Wild times"  
[X Link](https://x.com/HendrixLiu29208/status/1973124766052864496)  2025-09-30T20:36Z [---] followers, [---] engagements


"Huge congrats to the @mem0ai team on raising the $24M seed + Series A Super proud @keywordsai could help power the journey since the early days 💙 Memory is what makes us human. It's also what makes AI truly intelligent. @mem0ai has raised $24M to build the universal memory layer for AI. Thousands of teams in production. 14M downloads. 41K GitHub stars. Intelligence needs memory & we're building it for everyone. More👇 https://t.co/7r2zHCnNYh Memory is what makes us human. It's also what makes AI truly intelligent. @mem0ai has raised $24M to build the universal memory layer for AI. Thousands"  
[X Link](https://x.com/HendrixLiu29208/status/1983208507810685213)  2025-10-28T16:25Z [---] followers, [---] engagements


"We just crossed 10B @OpenAI tokens. Never thought wed reach this kind of traffic or get recognition from OpenAI More updates coming soon at @keywordsai"  
[X Link](https://x.com/HendrixLiu29208/status/1988348552536617165)  2025-11-11T20:50Z [---] followers, [---] engagements


"Huge mistake driving into SF at 7am today. Totally forgot how soul-sucking the commute actually is 💀 Sat there for an hour barely moving just watching my most productive time of the day evaporate At our hacker house my commute is literally [--] seconds. Bed ➡ Coffee ➡ Code Losing that morning focus hour to traffic feels like a massive L for a startup. Never taking "remote" for granted again"  
[X Link](https://x.com/HendrixLiu29208/status/1998822053340918024)  2025-12-10T18:28Z [---] followers, [---] engagements


"is the 'thinking' mode on chatgpt a placebo button i hit submit and it replies in [-----] seconds"  
[X Link](https://x.com/HendrixLiu29208/status/2013506743826084304)  2026-01-20T06:59Z [---] followers, [--] engagements


"Sulaiman was one of our earliest customers back in March [----]. At that time we had just joined YC and we were trying to build the first version of @keywordsai - a platform where developers could call LLMs in the same format and check LLM logs. I actually found Sulaiman under an X post where he was requesting a feature from one of our competitors but no one replied to him. So I reached out to him directly and we built out the feature he needed. Back then he was building a company named Invisibility with a very cool domain i(dot)inc which was basically a desktop version of ChatGPT. And until he"  
[X Link](https://x.com/HendrixLiu29208/status/2013731546184982870)  2026-01-20T21:53Z [---] followers, [---] engagements


"another YC founder just switched from langsmith to keywords ai. youre not getting founder-level support at 12am on a Saturday from most tools. founder support + product-led growth wins"  
[X Link](https://x.com/HendrixLiu29208/status/2015330931624821207)  2026-01-25T07:48Z [---] followers, [---] engagements


"its valentines weekend and my cto is on a date with claude code at a bar"  
[X Link](https://x.com/anyuser/status/2022565582256923003)  2026-02-14T06:56Z [---] followers, [---] engagements


"@owengretzinger @dominikkoch @langfuse @braintrust @usenotra curious if youve tried @keywordsai would love to get your take"  
[X Link](https://x.com/HendrixLiu29208/status/2022825662810329477)  2026-02-15T00:09Z [---] followers, [--] engagements


"@adisingh @agentmail looks so cool"  
[X Link](https://x.com/HendrixLiu29208/status/2015954626189918494)  2026-01-27T01:06Z [---] followers, [---] engagements


"One tiny prompt tweak - noticeably better LLM outputs. No reasoning prompt. No extra latency. No format changes. Just repeat the user query one more time. A recent Google Research paper found prompt duplication boosts accuracy on non-reasoning tasks across GPT Gemini Claude and DeepSeek - same cost better results. It wins 47/70 benchmarks with zero losses. Why it might work (intuition): LLMs are causal + order-sensitive. Repeating the prompt gives the model a second clean pass to lock onto the full context before answering. Practical tip: For classification extraction routing and other dont"  
[X Link](https://x.com/anyuser/status/2020985803133034699)  2026-02-09T22:18Z [---] followers, [---] engagements


"If you want an easy way to test prompt changes and see real results please check out @keywordsai Here's the article link: https://arxiv.org/pdf/2512.14982 https://arxiv.org/pdf/2512.14982"  
[X Link](https://x.com/anyuser/status/2020986269598441511)  2026-02-09T22:20Z [---] followers, [---] engagements


"milestone +1: first 6-figure deal just hit the bank proud of the team - just getting started"  
[X Link](https://x.com/anyuser/status/2021659902155141229)  2026-02-11T18:57Z [---] followers, [----] engagements


"customer in slack: Id be so grateful - Ill buy dinner for the whole house. he wanted a button to copy the llm convo from json (no base64). we shipped it in [--] hour. later: uber rings. huge chick-fil-a at our door. from him"  
[X Link](https://x.com/anyuser/status/2022032235147518406)  2026-02-12T19:37Z [---] followers, [---] engagements


"Open source LLMs are great but they can be painful to test and integrate before you ship. Today we shipped custom LLM providers in @keywordsai You can add any provider (just a base URL) register any model and use it across your workflow: - run evals before deployment - iterate on prompts - route it through our LLM gateway like your other models If youre using open source models in production and want to try this check keywordsai dot co. https://twitter.com/i/web/status/2011271544409112755 https://twitter.com/i/web/status/2011271544409112755"  
[X Link](https://x.com/HendrixLiu29208/status/2011271544409112755)  2026-01-14T02:57Z [---] followers, [---] engagements


"Were seeing more developers use open-source LLMs on Keywords AI - even as proprietary models keep getting better. Why - Control: keep data in your own infra - Cost: cheaper + faster at scale - Flexibility: customize models and avoid lock-in So I wrote the [----] guide to the best open-source LLMs broken down by use case. Read the blog 👇 https://twitter.com/i/web/status/2011525717159473213 https://twitter.com/i/web/status/2011525717159473213"  
[X Link](https://x.com/HendrixLiu29208/status/2011525717159473213)  2026-01-14T19:47Z [---] followers, [---] engagements


"you only see this in sf @HyperboundAI"  
[X Link](https://x.com/anyuser/status/2022463923115692403)  2026-02-14T00:12Z [---] followers, [---] engagements


"@owengretzinger @dominikkoch @langfuse @braintrust @usenotra @keywordsai sg"  
[X Link](https://x.com/HendrixLiu29208/status/2022826719250616676)  2026-02-15T00:14Z [---] followers, [--] engagements


"Two years after graduating I went back to UIUC and spoke to hundreds of students with my co-founder Andy and YC partner Pete. Still feels unreal. Not long ago I was sitting in the audience - now I was on stage sharing our story. We started with $0 built while in school went through YC survived [--] years on $500K and became ramen profitable. Most grateful part: I get to build with my best friends/college roommates and now Keywords AI is used by hundreds of AI teams in real products. Illinois still feels like home. Cant wait to be back"  
[X Link](https://x.com/anyuser/status/2018783685420695598)  2026-02-03T20:28Z [---] followers, [----] engagements


"RT @Andydy42: We did YC W24 survived on $500k and still havent burned through it. Were now a team of [--]. If $500k cant get you to rea"  
[X Link](https://x.com/anyuser/status/2023151369201525192)  2026-02-15T21:44Z [---] followers, [--] engagements


"We did YC W24 survived on $500k and still havent burned through it. Were now a team of [--]. If $500k cant get you to real signal the problem isnt capital. Unpopular opinion but $500K pre-seed rounds are plenty for most startups. Teams nowadays are staying at [--] members for the first year are hyper efficient with AI agents and generating revenue quickly. Why do you need $2M if youre not doing hard tech or deep research Unpopular opinion but $500K pre-seed rounds are plenty for most startups. Teams nowadays are staying at [--] members for the first year are hyper efficient with AI agents and"  
[X Link](https://x.com/anyuser/status/2022876408918540394)  2026-02-15T03:31Z [---] followers, 55.1K engagements


"Unpopular opinion but $500K pre-seed rounds are plenty for most startups. Teams nowadays are staying at [--] members for the first year are hyper efficient with AI agents and generating revenue quickly. Why do you need $2M if youre not doing hard tech or deep research"  
[X Link](https://x.com/anyuser/status/2022366171400720431)  2026-02-13T17:44Z 48K followers, 131K engagements


"Almost every YC founder Ive talked to switched from Cursor to Claude Code. Am I the only one still on Cursor"  
[X Link](https://x.com/anyuser/status/2022577439399157891)  2026-02-14T07:43Z [---] followers, 124.6K engagements


"its valentines weekend and my cto is on a date with claude code at a bar"  
[X Link](https://x.com/anyuser/status/2022565582256923003)  2026-02-14T06:56Z [---] followers, [---] engagements


"About a month ago @HendrixLiu29208 told me I should post a market map for AI agents. I agreed in theory 😅 But I kept delaying it. Because I wasnt satisfied with the tool choices. Last summer I posted an AI market map on Reddit and it got a surprisingly strong response. Since then the space has changed a lot. Over the past six months especially I kept coming back to one question: What actually changed and what truly matters now After a long time debating removing re-adding and second-guessing I finally landed on the first version of this market map. This time I chose to focus only on AI"  
[X Link](https://x.com/anyuser/status/2022478854246535517)  2026-02-14T01:11Z [--] followers, [---] engagements


"you only see this in sf @HyperboundAI"  
[X Link](https://x.com/anyuser/status/2022463923115692403)  2026-02-14T00:12Z [---] followers, [---] engagements


"after a few months of paying for both gemini and chatgpt i canceled gemini today. i use these tools like a personal assistant: help me think learn and make decisions. for that memory matters more than anything else. with gemini i kept having to restate what im working on. id open a new chat and it felt like starting from zero again. chatgpt just remembers stuff automatically. over time it builds context about what im doing how i think and what i care about. that changes the whole experience. model-wise i mostly use the fast option. geminis fast model wasnt as strong for my needs and the pro"  
[X Link](https://x.com/anyuser/status/2022406383359099130)  2026-02-13T20:23Z [---] followers, [---] engagements


"customer in slack: Id be so grateful - Ill buy dinner for the whole house. he wanted a button to copy the llm convo from json (no base64). we shipped it in [--] hour. later: uber rings. huge chick-fil-a at our door. from him"  
[X Link](https://x.com/anyuser/status/2022032235147518406)  2026-02-12T19:37Z [---] followers, [---] engagements


".@tensol_ai turns OpenClaw into full-time AI employees for your company. They handle repetitive workflows across support engineering sales and more running 24/7 in a secure environment connected to your tools with full context of your business. Congrats on the launch @pratik_satija and @olivieropinotti https://www.ycombinator.com/launches/PQ9-tensol-ai-employees-for-your-company-built-on-openclaw https://www.ycombinator.com/launches/PQ9-tensol-ai-employees-for-your-company-built-on-openclaw"  
[X Link](https://x.com/anyuser/status/2022022899180114020)  2026-02-12T19:00Z 1.5M followers, 272.6K engagements


"milestone +1: first 6-figure deal just hit the bank proud of the team - just getting started"  
[X Link](https://x.com/anyuser/status/2021659902155141229)  2026-02-11T18:57Z [---] followers, [----] engagements


"Hiring tip I didnt expect: Hosting a hackathon giving referral bonuses. We went back to UIUC last month (YC talk + biggest AI hackathon on campus). Met a bunch of students with real Im building every day energy. We ran office hours 50+ chats extended several intern + full-time offers. Total cost: $3k. Cheaper than a referral bonus and way more efficient"  
[X Link](https://x.com/anyuser/status/2021335279328768310)  2026-02-10T21:27Z [---] followers, [---] engagements


"7 of us live in a hacker house in Alameda - $7.1k/mo total for living + office ($1k per person). very quiet and safe. honestly great for a small team heads-down building. SF rent is now $3K/month/room 🤯 Advith Robert and I need to move. Our current place is $1.8K/person and every other building in SoMa is either booked out or going to cost us $3-4K/person. My hypothesis is this is because so many AI companies are fully in-person in SF again SF rent is now $3K/month/room 🤯 Advith Robert and I need to move. Our current place is $1.8K/person and every other building in SoMa is either booked"  
[X Link](https://x.com/anyuser/status/2021031085027754172)  2026-02-10T01:18Z [---] followers, [---] engagements


"SF rent is now $3K/month/room 🤯 Advith Robert and I need to move. Our current place is $1.8K/person and every other building in SoMa is either booked out or going to cost us $3-4K/person. My hypothesis is this is because so many AI companies are fully in-person in SF again (like us). https://twitter.com/i/web/status/2020543143939105108 https://twitter.com/i/web/status/2020543143939105108"  
[X Link](https://x.com/anyuser/status/2020543143939105108)  2026-02-08T17:00Z [----] followers, 251.2K engagements


"YC apps in 2025: cursor for x YC apps in 2026: openclaw for x"  
[X Link](https://x.com/anyuser/status/2021029640832417811)  2026-02-10T01:13Z [---] followers, [---] engagements


"One tiny prompt tweak - noticeably better LLM outputs. No reasoning prompt. No extra latency. No format changes. Just repeat the user query one more time. A recent Google Research paper found prompt duplication boosts accuracy on non-reasoning tasks across GPT Gemini Claude and DeepSeek - same cost better results. It wins 47/70 benchmarks with zero losses. Why it might work (intuition): LLMs are causal + order-sensitive. Repeating the prompt gives the model a second clean pass to lock onto the full context before answering. Practical tip: For classification extraction routing and other dont"  
[X Link](https://x.com/anyuser/status/2020985803133034699)  2026-02-09T22:18Z [---] followers, [---] engagements


"If you want an easy way to test prompt changes and see real results please check out @keywordsai Here's the article link: https://arxiv.org/pdf/2512.14982 https://arxiv.org/pdf/2512.14982"  
[X Link](https://x.com/anyuser/status/2020986269598441511)  2026-02-09T22:20Z [---] followers, [---] engagements


"Just one month after graduating I hosted the largest AI hackathon at UIUC by @keywordsai Huge thanks to #buildillinois and #FoundersUIUC for the support. I was blown away by the quality of demos. Some teams produced videos and products that could genuinely compete with startup demos that cost thousands of dollars to produce. There were also moments that felt especially surreal seeing people build ideas Ive personally experimented with and push them much further. Massive congrats to our final winners. Its incredibly impressive what they shipped in such a short time: Harvest AI turns real-world"  
[X Link](https://x.com/anyuser/status/2018425091898372254)  2026-02-02T20:43Z [--] followers, [---] engagements


"in nyc this week (mon - sun) would love to grab coffee and chat with startup founders. dm if youre around ☕"  
[X Link](https://x.com/anyuser/status/2018373372724277692)  2026-02-02T17:18Z [---] followers, [---] engagements


"Sweep 0.5B is out It beats 7B models in autocomplete performance while being 14x smaller. https://huggingface.co/sweepai/sweep-next-edit-0.5B https://huggingface.co/sweepai/sweep-next-edit-0.5B"  
[X Link](https://x.com/anyuser/status/2016586787632939481)  2026-01-28T18:58Z [----] followers, 26.3K engagements


"we just shipped a big improvement to how you rerun LLM logs. a lot of our users love bringing production logs into our prompt playground to tweak prompts and compare models. today we made that flow faster. now you can click any log and rerun it right in the side panel. you can try different models and run as many attempts as you want. you can also continue the conversation by typing in the bottom of the panel. this makes it easy to test and iterate on real production logs in seconds. check it out at @keywordsai"  
[X Link](https://x.com/anyuser/status/2016576224693514481)  2026-01-28T18:16Z [---] followers, [---] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing