Thats a good tip. I guess the question that brings up is solving the problem of discovering those communities that are still alive and thriving out there.
A friend of mine and I have been talking about ways to consolidate and improve online commenting. If you look up my profile and contact me about it I'd be glad to talk about it.
We provide support for optimistic updates and offline mode out of the box. Our idea is to give you the best of both worlds in terms of Firebase and Supabase. [1]
Thanks for trying it out! Right now we give the game engine some initial examples, and some character details are stored outside of the LLM context.
The token limit problem / memory isn't really being handled very well yet.
There is some character consistency slippage that does happen with the current mechanism as well so hoping to tackle both of these in next version.
Trying to figure out the best way to get community input, long term would be great if anyone can make a custom RPG with custom world rules and lore.
yea, AI Dungeon is OK but would love to see more here. Have some workarounds for memory if you are using langchain and happy to share. Feel free to shoot me a message on Discord geembop0x#9165
Most successful businesses in the U.S. are not technology companies at all, they just solve problems in creative ways. Look at someone like Wayne Huizenga who started AutoNation, Blockbuster, and Waste Management. That is like if Computer/Math prodigy Reed Hastings started a couple billion dollar companies in the Auto and Waste industry alongside Netflix.
The point being that great entrepreneurs can repeatedly find problems to solve and create value from.
For example, the author of this article was a successful entrepreneur in his own right. "Previously, the Chief Executive Officer of The Tie Bar - the #1 e-commerce destination for stylish men's accessories. Prior to The Tie Bar, Alter was a co-founder and President/CEO of SurePayroll, a SaaS technology company that is now a wholly owned subsidiary of Paychex®. Alter co-founded SurePayroll in 2000 after six years with McKinsey and Company, where he was a co-founder and leader of its Service Operations Practice."
The mention of Steve Blank's methodology is interesting. The best work I've seen done on this phenomena of customer-value-driven entrepreneurship is from https://www.effectuation.org/....basically that entrepreneurs start with what they have and the problems they know and go from there.
The counterargument is the Steve Jobs/Mark Zuckerburg argument for having a concrete vision and stopping at nothing to achieve a goal. But let's not forget that Apple started selling the Apple I which was primarily for hobbyists, and the Apple II didn't blow up until spreadsheets came out. Zuck was about to dedicate as many resources to Wirehog as he was to Facebook in the early days. Not to mention YouTube started as a dating site and many more classic examples of entrepreneurs having to figure it out as they go along.
I wonder what Keith Rabois would have to say to this. He has an interesting perspective in that he thinks founders will their companies into existence with clear visions that are often unwavering, and doesn't believe in the "lean" approach to building companies. Interestingly enough, Khosla where Keith worked earlier is the namesake of Vinod Khosla who does like the Effectual model of entrepreneurship (which is basically "lean" customer development based entrepreneurship). Another interesting thing to note is that most academic research into the age-old management question of what makes a good entrepreneur has far more to do with an individual's unique approach and perspective in the world, than it has to do with technology. Very few researchers are looking into "technology driven entrepreneurship" as much as they are personality, disposition, persuasion techniques, idea generation techniques, macroeconomic factors. This might change, but if anything when researchers look at entrepreneurial skill, technology skills are not the main focus. Again, this may be changing.
Personally, I think the only type of entrepreneurship that makes sense is very bottoms up. Look at the U-Haul story. The founder of that company simply did things none of his competitors would do because they thought it too risky.
My VC professor was the other co-founder of SurePayroll. He and Alter actually presented the idea to the CEO at Paychex in the late 90's. The CEO told them it was an extremely stupid idea that would never take.
About a decade later that same CEO was writing a check to them for $100M+.
They didn't even know how to design the algorithm for calculating payroll. They put the site up and had some elderly ladies who loved doing payroll calculate everything by hand and then send it back to a user within 24 hours. This went on for two months until they had figured out the algorithm.
In a world where consumer spend is going to be shrinking amidst an upcoming downturn and consumer CAC going through the rough "vertical enterprise software" is going to be the next megatrend.
Sabre does 3.87B in revenue and Duffel is just re-inventing Sabre with a better suite of APIs. What other legacy enterprise, vertical software solutions exist that people could "Duffel?"
> Sabre does 3.87B in revenue and Duffel is just re-inventing Sabre with a better suite of APIs. What other legacy enterprise, vertical software solutions exist that people could "Duffel?"
I could enumerate 100's of these problems, but this really isn't the right question to ask. The only reason this startup is possible is because the IATA allowed them to be. So perhaps the question should be "how many industries are ready to standardize their communication protocols via modern APIs?"
Great example - providing a unified banking API is clearly a multi-billion dollar a year business. But it requires the network effect of all of the banks to be involved. This either happens via industry groups or via regulation.
This is the essential issue with the travel industry. The data, and especially integration with the GDS's themselves is tightly controlled.
I've had many an idea for travel startups but data has been a non-starter.
I give Duffel props for having the mojo/influence/connections/funding to successfully implement something like this in an industry that has had few players. The only one I can think of is ITA but they are no longer independent.
From what I understand, Duffel are connecting to NDC APIs exposed by the airlines (or rather, by the Airlines' IT provider, i.e. Sabre and Amadeus... but that's another story). And on top of that, they expose a nice REST API (that is very similar in philosophy and structure to the NDC XML - Offer, Order). This is a win for the airlines, because every booking through Duffel is a booking without paying the GDS tax
1. Most big banks require a ton of red tape to use those APIs. Even if you're using modern technology with a top tier dev team, this is still a 4-6 week endeavor at its absolute quickest turnaround.
2. All ACH transfers, for example, have to clear the Fed, which is done via FTP.
3. The consumer use cases are endless (make payments, analytics, etc). Just look at Mint, Truebill, etc.
4. The commercial user cases are endless (treasury management, AP, AR, etc).
(Disclosure: I'm the CEO of Duffel) Totally agree with your point above re: IATA. They made it possible but for a different reason, they opened the door to competition.
Standardisation of communication protocol can only do so much, especially when the standard is built to accommodate the needs of 100s of airlines using 100s of different host systems.
Also, there is ton of red tape around airlines' APIs as well, I'd say probably as much as in banking.
On the consumer banking side I would agree with you, not many individuals have complex enough banking needs where programmatic access would be a value-add. However if you look at businesses, any company doing more than (let's say) 50 bank-to-bank payments a month I think would benefit from being able to automate their bank account interactions in some fashion.
And I would humbly ask for someone to think again about the regular people on a more pragmatic or tech-savvy end. My banking is relatively simple, and yet many of my needs are not covered by any existing solutions. The absolute basics I want is an API access to my current per-account balance. Another step would be programmatic access to per-account transaction history.
I don't want to use banking apps and sites daily, they all offer garbage UX that's designed for upselling credit cards, consumer loans and other financial product. What I want is simple API access. I can take it from there, and do a better job than banks do, because my own incentives align better with my own interest.
I mean I feel like "need" is a bad choice, when the point is that it can be a lot better. I didn't need Venmo to send money to friends, but it is quite clearly better than pre-existing options. Also there are plenty of shitty banks out there with their own insular network and plenty of money they are working with. What options do you have in the Middle East per say?
You could probably say the same about Stripe when they just launched. Authorize.net was doing billions in revenue and stripe just re-invented them with a better suite of APIs.
Truepill is an API for pharmacies that is doing seriously well and is under-hyped. The API-ification of things is a serious boon for developers everywhere.
But financial services (payments) and travel are just two verticals. There are vertical software solutions in healthcare, other aspects of FinTech (lending/wealth management), retail that can probably be re-vamped.
The question is what other legacy providers other than Authorize.net/Sabre exist. Where do we find software solutions like this?
Stripe is only loosely synonymous with Authorize.net
Your missing the part where Stripe also bundled the payment processing (Stripe uses Vantiv AFAIK) and the merchant account (which you need to get from a Bank). Stripe's innovation was to simply flatten the stack across gateway, processor and merchant bank account.
Stripe has an incredible API, but it's only one half of the equation here.
The $1.4B Sabre does in airline bookings revenue isn't insignificant but don't underestimate the difficulty of reinventing Sabre with a better suite of APIs from both a technical and commercial standpoint though.
I think access to bank APIs is a similar space. Companies like Tink or Truelayer used to rely on credential passing but now with the wake of open bank apis (PTSD2) they are implementing standard oauth flows. That being said, I think that the biggest entry-level problem in that space is an enormous level of regulations regarding bank data. You can't even start using their API without your app being validated by them.
Isn't this the space that Plaid is taking the lead in? What other opportunities are there in the banking API system aside from what Plaid is doing/won't do in the near future?
The glasses industry is so structurally bundled and unsound that a disruptor like Warby Parker was able to come in and build a billion dollar business in a matter of years.
Hearing aids are similar as mentioned on the thread.
Wondering what other industries have these structurally integrated industries where a disruptor could deliver a tremendous amount of value with insider Knowledge or reintegrating a supply chain. I believe an original Warby team member worked in the eyewear industry.
I have a feeling things like Car Dealerships are next. You fundamentally have to be able to drive massive value to consumers by making a one time purchase expneisve good cheaper with an innovative business model.
reply