Huge update.. I joined @Cognition I’m so incredibly proud of what we built at @PromptLayer. Although it feels a little surreal, I know the team is left in amazing hands. I’ll be staying on as an advisor and can’t wait to see where they take it. We started PromptLayer four years ago, and we practically coined the term “prompt engineering”. ChatGPT had come out weeks prior. Nobody knew exactly how LLMs would evolve, but we did know it will completely change the way we build. We were the first developer platform for this new type of builder. Cognition feels like an extension of this vision. Devin is how this new builder will imagine, invent, and create. We're entering the age of software abundance, to steal a term from @scottwu46 Everything is Coding Agents. It’s the single biggest problem in AI. Cognition is one of the fastest growing companies ever. But I think the team is what really sold it for me- it's packed with so many former founders Excited to share more soon...

Engineer friend just quit FB to start a startup Planning to do a dating app I cannot interfere this is a canon event

i would like to report a murder

Found myself wishing for a devtool while hacking on some @OpenAI GPT-3 projects... Introducing PromptLayer (promptlayer.com), the first ever platform for prompt engineering. Totally free. Track, debug, and tag old completions. Build prompts thru trial & exploration.

What is going to boom about Lower Manhattan now that all the people are gone? Congestion may be effective, but not in the way you want.

Same exact energy. Kol HaKavod.

exited founders returning to nyc after achieving glory in gaul (sf)

honestly this would be lit

Hilarious professor. crushed some brews with him after the CS189 final when I was in school

Had a great conversation with the one and only @stephen_wolfram 🗣 We talked about prompt engineering, AI ethics & constitutions, and of course the brand new ChatGPT plugins @WolframResearch + @OpenAI collaboration is pretty historic. piped.video/watch?v=xKV8_N1b…

Don't learn to code, learn to prompt engineer. Excited to announce our $4.8M raise led by @neversupervised at ScOp, @pboyce at Stellation Capital, along with @joshuabrowder, @gokulr, @benln, @romainhuet, @OfficialLoganK, @ByrneHobart, @akilian,@GabeStengel and operators from OpenAI, Anthropic, Microsoft, and Google. @Jonpon101 and I built PromptLayer for domain experts like doctors, lawyers, and therapists— not for ML engineers. Today, hundreds of the top AI teams use our prompt CMS for prompt management, evals, and observability. techcrunch.com/2025/02/07/pr… Come join us @promptlayer - we are hiring in NYC 🗽

Born to Run by @springsteen turns 40 today... should be a national holiday. theatlantic.com/entertainmen…

gpt-4o blows gpt-4-turbo out of the water. So quick & seemingly better answer. Also love the split-screen playground view from @OpenAI

And to make it even cooler... you can read his thought process in realtime. Had the idea of using Wolfram + ChatGPT two months ago and now here we are writings.stephenwolfram.com/…

Today, we are releasing a product for the most important part of GPT apps... the prompts. Introducing PromptRegistry by @promptlayer 🍰 Access & publish prompts through an API 🍰 Build A/B testing 🍰 Version control Get those damned prompts out of your prod codebase 👊

This is CRAZY!! Using @OpenAI's new Operator agent to use our product. This is really unbelievable. I asked it to audit our platform. It created an account, read through instructions, tried things, fixed issues, and figured it out. Just wow.

if there’s anytime to bring back this fit it’s now

The biggest unknown in AI today is 📝 evaluations 📝 We spoke to 30+ teams using LLMs in production and this is the top challenge. 🤨 Does PromptB work better than PromptA? 🤨 Will it cost more? 🤨 Will it break my product? Today’s @promptlayer release helps 🍰 learnings ⤵️

if you're launching with a Forgot Password button you're launching too late!

NYC AI scene is so underrated. @promptlayer event is epic 🍰

New @promptlayer HQ 🍰 📍 Soho, NYC 🍰 Extra desks for visiting hackers

user API spend on @promptlayer @OpenAI 📈📈📈

on my way to @OpenAI DevDay! find me for a hat I got a few extra 🍰

Funny story. Even though Natasha just covered @promptlayer's seed round announcement, we actually met for the first time over a decade ago. She interviewed me after winning the TechCrunch hackathon when I was only 16. I won't link the interview because it is severely embarrassing... but it was great to reconnect at this next stage of life 🙃

Just saw the king of NYC himself while eating lunch at the park. Jerry Seinfeld. The legend. I know every episode by heart.. starstruck but I had to ask for a picture. "Hey! Do you mind if I get a quick selfie?" "No, sorry I'm with my family." Getting rejected by Jerry might actually just be the quintessential Jerry Seinfeld experience. It was amazing. Plus.. now AI is basically good enough to generate a selfie.

So sick. Just tried it out with @langchain + @promptlayer

I call it Persia

don’t wear headphones at the gym look up from your phone on the subway let yourself get a little bored break through the dystopia

Happy birthday to @PromptLayer 🍰🎉🦾 We launched v1 exactly one year ago. It started as just a dead simple @OpenAI logging tool. Since then we added the Prompt Registry, evaluations, playground, fine-tuning, multi-modal support ... and much much more 🏋️‍♂️ See you in year 2.

only 2 dollars for repeat offenders! not bad :)

Context engineering continues to pick up steam Still might be a fad or blip... too soon to tell

I love Cursor. They are such a great example of the crawl, walk, run process of software. First version was pretty jenky. It's gotten so much more polished since then. Can't live without it. Don't wait to ship!

imortant skill. I think of it less like avoiding questions and more like steering conversation where you want it to go. trump-style interviews

Money laundering. One cartel to another

How @tryramp does agents at scale. Great talk by the great @rahulgs of jsonformer fame 🫡

Check out @SliceCapital! They just launched on @ProductHunt. Making it easy for normal people to invest in startups producthunt.com/upcoming/sli…

The next phase in @promptlayer’s master plan.. 🔭 User event & agent analytics 🔭 Every request can now be tagged with user behavior & agent metadata. Stop jerry-rigging Mixpanel & Segment in your LLM stack. Commonly requested feature from the PMs and SREs using PromptLayer.

Who's the #1 Prompt Engineer in NYC? We're hosting the 1st ever Prompt Engineering Tournament 🏆 Compete in 3 rounds, win a @rabbit_hmi. Win the glory of your peers. Learn some skills. Someone will be crowned 👑 lu.ma/cs8vn7es

🍰⛓ Big news. You can now use @promptlayer with @langchain Langchain speeds up development by 10x. Promptlayer speeds up debugging by 10x. Start using them together with one line.

It's like wheels on suitcases. So obvious but took a while to get there

Happy Hanukkah from the @promptlayer HQ! 🕎

The best narratives are unprovable. And the best narrators choose them for this reason Then turn it up, exaggerate, and lean in even harder It's the way of the world and showmanship Keep saying something until everyone believes it (including yourself)

You can now opt-in to open source your completions data in @promptlayer 🍰 Free and open LLMs like GPT4All are the future 🗽

Welcome to the team @bucky_roberts 🍰 Huge things coming. Be ready.

"I'm not a vibe-coder, I'm way worse" @iamjasonlevin

Is GPT-4 responding slow today? 🐌 Couldn't tell if my code is slow... or if @OpenAI is lagging... So we built a graph to show the average latency of @promptlayer users for 3.5-turbo, 3, and GPT-4 📊 Check it out at promptlayer.com/latency

Often find myself going back to this diagram of the prompt engineering lifecycle. ⚡️🌀 LLM iteration is zigzag, not linear. There is no "correct" way. It's all just tinkering. (From a talk at @OpenAI DevDay by @colintjarvis and @john__allard)

crawl walk run

Something just tells me @promptlayer is going to be the biggest thing I do in my life…. we’re just getting started 👀👀👀

The best prompt engineers I've met are nontechnical 🏆 ... but everyone is focusing on the PhD ML engineers. Launching a new visual evaluations feature on @promptlayer 🍰 Designed for product teams. Stop gatekeeping the cool stuff— build as a team.

Israel must be the first country to notably combine cyber attacks with kinetic attacks. Stuxnet is so 2010.

Everything in life is prompt engineering. Made these hats- can I send you one?

openai can release a model that just pretends to be thinking for 30min and so many of you would pay extra and say it’s agi

Psssssst. PromptLayer is now natively integrated into Langchain

New yorkers are not bothered by honking. Yes, you can solve traffic by removing the people. But that's not a good thing either. What about the family who wants to take their kids into the city for a show? Tell them to wear fire retardant jackets on the subway in

8 new Anthropic prompt engineering tips I learned from @zswitten: ​ Tip #1: Use XML tags to structure prompts Tip #2: Be specific with instructions rather than saying "be concise" Tip #3: Mimic desired tone of response in your prompt Tip #4: Proofreading your prompts matters Tip #5: Write prompts in a language you know well Tip #6: Use Claude prefills to force JSON response Tip #7: Use negative prompting sparingly Tip #8: Rely more on human messages than system messages ​ Thanks to Anthropic's "Prompt Doctor" for the great workshop at AI Engineer 🙏

Now with zero cost to them (their loans will be payed back), great moves by @sama and @vkhosla. Already valley legends but even more proof

Be ready… @promptlayer x @honeyhiveai 🍰🐝

i put my card down at dinner with 20 friends... dreading the "why did you venmo request so much!?" texts 2023 mode is sending chatgpt links 👊

Really fun conversation Hot takes about AGI… and of course prompt engineering Thanks @danshipper

It's time to believe in miracles again.

being a founder is just resisting the urge to spend all day every day hacking. that's dessert. you gotta eat your veggies too 🥦 [talking to customers]

The community around prompt engineering is unbelievable 🚀 LangChainHub is a community-curated collection of the most useful prompts. Pretty much all you need. The @promptlayer Registry comes with it. Visualize and start working with these @langchain prompts in seconds.

This was a fun interview. Talking prompts, humanity, and how we are building the future of AI work. But I will say @iamjasonlevin is a wild boy..

עם ישראל חי 🇮🇱

have zero regrets about making the @promptlayer logo an emoji 🍰

Top floor of @moodysanalytics speaking about prompt engineering and giving a @promptlayer demo 🍰 thanks for having me at Data Driven NYC @firstmarkcap @mattturck

bravest thing you can do is send a joke to a 10+ person groupchat. i'll always throw a laughing emoji even if it's not that funny. i stand with the brave & courageous.

Kintsugi 金継ぎ is the Japanese practice of using gold to repair broken pottery. Literally translates to "golden joinery". An ancient art that feels like a great metaphor for @nntaleb's concept of antifragility. Pottery is made even more beautiful through breakage.

You can now use PromptLayer to track the new ChatGPT API with @langchain 🍰🔗

Starting the week off with a product launch 🍰 shiny new @promptlayer analytics dashboard!

The @promptlayer team will be at the @langchain NYC meetup tonight with some highly limited edition stickers 🍰 Come say hi to me and @Jonpon101 👋

my talk at data driven nyc just went live 🍰 piped.video/watch?v=YU-A2SNv… (we're hiring for eng roles — dm me!) thanks again to @mattturck @FirstMarkCap for hosting

new @promptlayer job posting front end product engineer DM me if interested or know someone 🍰

Building a startup is about dialectics. Extreme urgency / 10-year gameplan Obsession with details / imagation of the future Insatiable work ethic / celebrating wins PromptLayer NY team went out to celebrate a revenue milestone! Steak, some drinks, then back to working on the next goal 🥩 🍰 Acknowledge wins, but don’t rest

these days anytime I need to do something complicated like changing the borders of a button

Demos at the Generative AI NYC hackathon have begun!! Favorites below ⬇️

SO EXCITED

test-driven prompt engineering. this is way too fun 🍰

I saw Soham Parekh at the Ramp HQ yesterday. I told him how cool it was to meet him in person, but I didn’t want to be a douche and bother him and ask him for photos or anything. He said, “Oh, like you’re doing now?” I was taken aback, and all I could say was “Huh?” but he kept cutting me off and going “huh? huh? huh?” while snapping his hand shut inches from my face. I walked away and finished my shopping, and I heard him chuckle as I left. When I went to pay, I saw him trying to walk out the doors with about fifteen company Apple Vision Pros in his arms without paying. The cashier was totally professional and said, “Sir, you need to pay for those first.” At first he pretended to be tired and ignored her, but eventually he turned back, dropped them on the counter, and let her scan them. When she started ringing one Apple Vision Pro several times, he stopped her and told her to scan each one individually “to prevent any electrical interference,” then turned, winked at me, and yawned loudly every time she tried to tell him the total. I don’t even think “interference” is a real word. After she bagged them and named the price, he just kept cutting her off with huge yawns. Then he paid and strutted out.

The duo I know we've all been waiting for... Using @promptlayer with @replit is easy 🤝 Check out the template at replit.com/@bardia/PromptLay… Thanks to @thepericulum for putting it all together 🔥

Why fine-tuning is (probably) not for you 👀 ​ With the help of @pranav_kanchi, we spent some time talking to customers & researching the state of fine-tuning ​ Fine-tuning can be really powerful, but is way overhyped (at least today). It makes building a good LLM application harder. ​ The #1 thing you should worry about is how fast you can iterate on prompts & the AI application. Fine-tuning is slow. Training a new model takes time and is expensive. ​ In today's world of LLMs, you want to just be trying tons of ideas and seeing what sticks. ​ 🙅 Why you should avoid fine-tuning 🙅 1.RAG is usually just way more effective 2.Complex process: dealing with data, pipelines, and training 3.Slow iteration cycles means mistakes are less forgiving 4.Tons of hidden costs: Constant updates & dev time to name a few. 5.Large data needs. 6.Potential data privacy issues. ​ These things might change. I think we are entering a world where everyone will be fine-tuning OSS models for their use-cases. We just aren't there yet. ​ Everything in prompt engineering is a tool 🪛 Obviously there are some things that fine-tuning is actually the best tool in your toolbox. ​ ✅ When fine-tuning is the answer ✅ 1.Tailoring specific output formats 2.Edits tone & writing style 3.Improves complex reasoning 4.Reduces token usage for complex prompts. 5.Can get GPT-4 level performance from GPT-3.5 ​ #5 is the simplest and most common use case. Just train a cheaper model on your old, expensive outputs! Easy. ​ So why not use both? Maybe... but don't start with fine-tuning. ​ Fine-tuning has a lot of mindshare but it's mostly just VC Twitter buzz. You should be focused on shipping the best product as quick as possible.

what an amazing demo 💰

PromptLayer makes this so easy. Filter exports by tag, metadata, search queries. Happy prompting 🍰

crazy that some people still don't know @render!

OpenAI's new open source model is basically as good as o3. Which was already my go-to model. Previously, the best open source models were all Chinese. American AI is back 🇺🇸

sf mode for the next week 🌁

I would pay way more than $8/mo to not see AI influencers in my feed

Wow. 618 high school signups for @PennApps this season. The HS Hacker movement has exploded.

dating app, party app, many such cases

My sleep routine works every time: - prompt engineering until 2am - cold plunge (while thinking about new PL ideas) - no screens after 3:30am - meditate and contemplate where ilya is for 30min - 8 hours of sleep (not 8sleep, normal bed) - 1pm wake up & slam a bagel with lox

Shout out @ncweaver for being absolutely spot on in his post-lecture crypto rants 4 years ago. Re: centralization, re-inventing banks, already solved problems in finance.

Demo time! It’s always hard to give a @promptlayer demo following a @morgantepell @gritdotio demo because everyones jaws are already dropped.. 🍰🐜 @llama_index @lastmileai_dev @MongoDB @_changxu @RezwanKh

Gave a talk to the NYU Entrepreneur-in-Residence club “HOW TO BECOME AN EXPERT PROMPT ENGINEERING” Amazing group of killer founders. The kids are gonna be alright 🙏 Thanks @saketh_kesiraju for setting it up 🫡🍰

It's true The best AI products are built by domain experts. Not just ML engineers. Really cool to see @promptlayer on America's oldest newspaper.

Hey Lizzy! I run a firearm club for young Jews in NYC. We have 200 members mostly first timers. We got to the same range you are at. DM me would love to have you join the next range day

Just pitched to Stephan Wolfram about @hackBCA, and he had already heard of it. Woah.

The first prompt engineering tournament was a success. Thanks to all that came. More pics coming soon but all I have to say now is HACKER CULTURE IS BACK That was so fun. Prompt engineering is all that matters. // @promptlayer