I was listening to
@lennysan
's podcast with
@Patticus
and they touched on the concept of anti-active usage products and why Profitwell is free to use.
Decided to go deeper on this and found a crazy interesting framework on the Profitwell blog.
Let's break it down 👇
Have been playing around with
@supabase_io
and it is absolutely magical ✨
Might have just found my new favourite way to spin up a DB and Auth in minutes
🚀 Here's a quick demo of Lookup for wk1 of
#nws2
.
✨ Import your data, ask questions, and get results instantly with the power of AI ✨
Please sign up and try Lookup with your own data!
LINK:
cc:
@_buildspace
@FarzaTV
Am I off localhost now?
📚 I recently read Fundraising by
@ryantakesoff
Straight to the point and packed with hard-earned experience. I definitely recommend giving this playbook a read if you're interested in raising funds for your business!
My 3 key takeaways 👇
@eriktorenberg
As someone who worked in finance and consulting straight out of uni before switching to tech.
Two main drivers:
1. Awareness - I didn't know about the great opps available in tech
2. Education - my perception was that I would be disadvantaged in tech given my lack of CS degree
Starting to think that the narrative around incumbents capturing most of the value in AI is wrong or at the very least lacking a lot of nuance.
As someone who has been building with LLMs and keeping up with the space, many of the underlying assumptions may not hold. Distribution
@carlvellotti
I've had success with using ChatGPT to brainstorm ideas. Getting it to respond in a table has been helpful in getting more options out of it. E.g. "Brainstorm ideas for X and assess them across criteria X, Y, Z"
It's also surprisingly good at scaffolding PRDs, which has been a
@jmt_nft
Agree with the sentiment, but slightly disagree with the analogy.
Amazon was one of many tech stocks.
Most are no longer around.
Picking Amazon vs the other tech stocks was the hard part.
Most NFTs will be worth 0 in 10 years time, some will be worth $$$$$
Just shipped something super cool in
@useLookup
Lookup will now suggest questions for you to ask based on the data you upload
Here's a 30s demo, lmk what you think
Finally got around to cleaning up the landing page for
It's still ugly af, but a material improvement from how bad it was before.
Next, I need a much slicker product demo video 😅
cc:
@_buildspace
@FarzaTV
It's a huge 🚩 if you're not already using ChatGPT in your job as a product manager.
Don't worry I'm here to help, this is a running thread of techniques, prompts, tricks and tactics for using ChatGPT as a product manager.
Let's get straight into it 👇
✨ Shipping something new to kick off the year ✨
🔎 Introducing data previews in Lookup!
Easily view columns and samples across all of your datasources. Super helpful for when you've imported lots of data and can't keep track.
Here's a 30s demo.
cc:
@_buildspace
@FarzaTV
Imported a csv dataset from Kaggle:
and started asking it questions.
What's super cool is that there's no column in the dataset called video title, yet it was smart enough to return the most similar column (video_desc) instead 🤯
🕐 Start Early
Build a network and plant the seeds way before you actually start raising.
Have informal conversations with investors, gauge interest, filter for fit, build regular touchpoints.
You want to know your raise will succeed before you even start.
🚀 Momentum is Everything
Fundraising is purely a matter of momentum.
Starting early ⬆️ means you can close rounds faster.
Business momentum is just as important. Don't play all your cards at once, stagger important milestones.
@_michaellin
“I realized what the real cost to golden handcuffs was. The cost is your youth, your time, and your life. People don’t accurately judge these costs, because a salary is a hard number, whereas the value of your youth is more intangible.”
This 💯
I’ve been quiet on Twitter the past few days as I’ve been working on some fundamental changes to the
@useLookup
execution engine.
Just tested it out and the results are blowing my mind 🤯
Literally bouncing with excitement - Can't wait to share it with everyone!
🔥 Launching Sponsorwatch - your free weekly newsletter for finding newsletter sponsors!
👀 We hunt down companies backing biz & tech newsletters every week, and deliver them to you in a neat list.
Dive into the first issue here:
🧘 Mindset
YOU are worthy. YOU are the scarce resource. Money is plentiful these days.
Be confident and kind.
Fundraising is intense. Look after your health before and during the process.
@eriktorenberg
To add to this.
1. Perception amongst (most) of my consulting/IBD friends is still very much that you will earn less in tech unless you are a SWE
2. Consulting does a much better job of selling "consultants come from all educational backgrounds"
Today I scraped the 130 plugins currently live on ChatGPT.
Some early findings:
- 90 of 130 are not using auth
- To paint a picture of maturity, most plugins with auth are using Bearer tokens (rather than oauth)
- Huge spread of industries and use cases covered including ...
@rafalwilinski
Are you actually getting GPT-4 level answers from the 70B model?
I'm loving the speed, but response quality isn't there yet in my experience (waiting for the bigger Llama3 model)
I recently tried
@usedouble_com
- A tool to help you
Clean, Enrich and Qualify leads using AI
This is one of those ideas where I can see a huge market for if the product is executed well and actually delivers on it’s marketing promise. Double is quite easy to pick up, and feels
Here's a thread of 50+ experts to follow on Twitter sorted by topic, including...
✅ Writing/Newsletter experts
✅ Productivity/Creativity experts
✅ Marketing experts
✅ Social Media experts
✅ Entrepreneurship experts
✅ Course experts
Here we go... 👇
With better outcome measurement, the fixation on active usage may no longer be valid.
This gives us two dimensions:
1. How measurable is your outcome and how much does your customer agree with your measurement?
2. How much can you truly automate?
I've been playing with different UI/UX patterns for AI.
Here's a fun one I just whipped up for Twitter.
Trigger reply suggestions by double-tapping the "?" key ✨
Some background, your product gives your customer an outcome that they seek.
But it can be hard to measure this "outcome".
Because we often can't measure it, we've grown accustomed to measuring a proxy - active usage. Basically, if customers are using it, it must be valuable.
There's a problem with this, usage doesn't necessarily correspond with an outcome or the best way to reach that outcome.
The example Patrick gives is a sales rep forced to input their progress into Salesforce. This data has no direct connection to a sale, it's just tracking it.
How does everyone evaluate the quality of their prompts?
Is there a tool that takes your prompt, subs in your data, sends to a LLM, then evaluates the completions based on specified criteria?
Might build this if it doesn’t exist
cc:
@Planana_ai
@_nwmo
@Shpigford
@yoheinakajima
Really starting to gravitate towards the concept of a university/school powered by Twitter threads.
Product seems like a great fit for this format
#buildinpublic
When I started building Lookup, I made sure to block out time to try other AI tools.
This helped me stay up to date on competitors, new UX patterns for AI, and build better AI products.
Given how useful it's been for me, I thought I should share. Starting today, I'll be
Customer feedback is critical for building a great product.
But analyzing customer emails is a hassle...unless you use AI 🤖
If your business uses a support@ or contact@ email as the main point of contact with customers, this step-by-step guide is for you.
Let's dive in! 👇
Optimizing for active usage was okay in the past, but customer expectations have changed.
We've moved from an age of "show your work" to "just do the work for me."
What if we just removed the usage? Why optimize for usage if we measure customer value directly?
This the coolest thing I've seen in AI all week.
The decoder level constraints look 👌🤯
Cannot wait to play around with this. Now if only they had a JS/TS package...
🚀 Excited to announce the first release of , a novel open source programming language and platform for language model interaction!
Combining prompts, constraints & scripting, LMQL elevates the capabilities of large language models.
🧵1/6 A quick tour.
Just shipped the
@useLookup
API 🚀
It's now easier than ever to start pushing all of your customer feedback into Lookup for automated analysis.
Stay tuned, tomorrow I'll share how we dogfood Lookup by using it to monitor it's own customer feedback! 🤯
Just shipped a pretty big "under the hood" change to Altero AI.
There's no major UI change, but you should start to see more reliable company metadata and logos flow through the platform over the next few days 🥳
Lesson learned here is that most company data API providers
Looks like it’s time to unveil my new project
We’re building out a network of creators open to sponsorships by providing free tooling, expert help, and pre-vetted connections. I’ll be applying all my learnings from
@thumblytics
and building in AI to give
I paid someone to do TikTok outreach for and he literally got ZERO results
Maybe need to hire agency next
It seems weirdly difficult to reach TikTokkers when you literally want to give them money
Maybe an opp for a biz?
I was interested in revenue concentration by country for a side-project, so I imported some Stripe data into Lookup and just asked.
It took just a few seconds to get the answer.
No data cleaning or anything like that - It's absolutely mindblowing how magical this is🤯
🚀 Here's a quick demo of Lookup for wk1 of
#nws2
.
✨ Import your data, ask questions, and get results instantly with the power of AI ✨
Please sign up and try Lookup with your own data!
LINK:
cc:
@_buildspace
@FarzaTV
Am I off localhost now?
Our AI agents can produce detailed business research on any company.
Put in any URL and get a research report in minutes.
Best of all, you get a comprehensive list of sources so you can check the AI agent's work or dive deeper yourself.
Early beta at:
Had a couple of hours tonight and was able to ship the email integration for
@useLookup
You can now forward customer emails, form responses, bugs + more directly into Lookup for AI analysis
Just generate an email and start forwarding
Here's a 60s video, lmk what you think!
✨ Announcement ✨
I'm launching a cohort-based course on Product Management.
I'll teach live classes, sharing everything I've learned about being a great product manager, with the goal of accelerating your early PM career.
Applications now open 👇
To build and ship a valuable product in a single day, I will need to focus on the smallest discrete piece of value I can deliver.
My guess as to the tradeoff of value vs time to implement for these 3 methods are:
manual > crowdsourced > automated/algorithm
Finally finished tagging 100+ wisdom-filled threads about product management from the master himself
@lennysan
Whose threads should I look at next?
#buildinpublic
Overall, it is super impressive what
@StephenAfamO
has been able to pull off. Swish has a clean and simple design, and seems to have all of the features one would expect from a blog / CMS platform.
Check it out:
It also interesting to note that these 3 methods are not mutually exclusive and can build on top of one another. For example, start with a manually curated set of tweet threads, start crowdsourcing votes and more tweet threads, use this data to kickstart automated curation.
Super interesting how different the sponsorship buying processes and dynamics are across different platforms.
E.g. Newsletter ads are often copy and paste vs YouTube sponsorships which often require extra effort to add into the video (or at least requires creator voiceover)
Currently studying up on the market research space and looking for companies to deep dive on.
On my list I have Crunchbase, Statista, Euromonitor, CapIQ, Bloomberg, plus a few more local ones. Any others I should add to my list?
How do we as product people leverage this?
1. Put measurable outcome into your product as much as possible (e.g. time/cost saved, revenue earned, etc.)
2. Focus on active usage only if it's in a workflow
I built something fun today ✨
A crowdsourced generative story:
1. Everyone can submit suggestions for the next chapter.
2. AI uses chooses a suggestion to generate a new chapter every day.
3. The story goes on forever.
Check it out 👉