đ building Tongue - an AI personal assistant
New York, NY
- Tweets 1,238
- Following 875
- Followers 338
- Likes 821
2) interior design. Complete with mocks of specific products in your space. 8/10 - I had to ask for it to go this direction / generate the images. It being slightly more proactive would be awesome. E.g âhey looks like youâre designing a space ⌠want me to generate 3 mocks in these stylesâ
1
23
3) monitor fb marketplace messages. Unfortunately 2/10 experience here. Took multiple prompts to get it to monitor all my listings, and yet another prompt to get it to monitor messenger (which it treats as different than fb marketplace inbox)
These sort of failure modes are distinct to ai - a human assistant would know to monitor new listings as well as anywhere messages show up. A human isnât like well messenger is a separate api - letâs ignore it.
The average consumer wants ai thatâs at or above human capacity and @Muse fell on its face here unfortunately.
28
Would be sick if @Muse persisted my home (and work location) or pulled it from Google Maps.
1
51
Too big to fail ⌠heâs talked about this.
At this debt level the bank wonât let you fail
'Rich Dad Poor Dad' self-help author Robert Kiyosaki is $1.2 billion in debt, per NYP
Community note
Kiyosaki does not hold $1.2 billion of debt. The debt is held collectively by a real estate investment partnership of which he is one member. Real estate investment groups routinely carry debt.
vanityfair.com/story/the-wildâŚ
52
Sending lots of love to the Apple community on my last day as CEO. My title changes tomorrow, but the love I have for the Apple community never will. Thank you for being a constant source of inspiration. My gratitude is endless, and Iâm excited for the next chapter!
10,594
26,811
318,066
28,705,267
Not really.
Bodies didnât decay - people just got the option to get fat (abundance)
Intelligence is now cheap. And you can think about higher order problems. (Abundance)
Feels like a weird anti ai take.
114
The exact problem with âdo it allâ assistants.
They need to perform at or better than a human across all vectors. And fail miserably in ways humans wonât.
Usually due to a lack of internal api for an ask - so the ai / harness literally canât do whatâs requested.
Iâm betting solving a niche and moving outwards is the way to go here.
AI products have this bizarre usage/learning curve that's the opposite of software products. Incredible magic up front, that deteriorates into annoying failures after the honeymoon period
103
Yeah - most businesses literally donât know how to make more money.
Everyone pretends they do. Whatâs your PM gonna say. I canât grow this product faster than what weâre getting from our network effect 20% yoy growth.
this isn't a token efficiency problem - it's the equivalent of trying to squeeze the last drops out of already-squeezed lemon.
companies that have been around for years already have a product with market fit (generally).
AI can make parts of the business faster, tackle automation, etc - but it won't magically 10x the entire company's productivity or revenue unless they're using it to build entirely new products.
this is totally unsurprising.
1
1
124
I think the reason is people have different use cases they want a personal ai agent to do. And most companies have (incorrectly) positioned themselves as a personal assistant that can do everything.
And invariably they canât.
I think the winner is going to niche down, build an agent that does one thing exceedingly well and slowly add use cases over time. Until they can credibly say this does 99% of the tasks people expect an ai personal assistant to do.
the reason that nobody is using agents is because they are still wildly unrealiable even including the wildly expensive frontier models that are far too expensive to put in a mass-market consumer product
the reason it works for coding is because you have a six-figure-salary subject-matter-expert resource sitting there babysitting it all day (and still shipping slop)
1
118
Chad Newbry retweeted
One misleading thing large app builders say is
"Got 100 million views all through organic content no ads"
Meanwhile they're running a 50 creator UGC campaign
Your PAID ugc campaign is not ORGANIC
When I was a beginner this disoriented me so much since I thought just posting organically everyday could make me rich
20
1
75
5,002
Good point. These use cases seem obvious but worth highlighting.
Too many people trying to get AI to write whole codebase. Not enough using it for the boring shit nobody has time for: Improve your error messages. Build better on call dashboards. Fix the broken stack traces in Sentry. Clean up your old feature flags. Low stakes stuff AI is actually good at and you weren't going to get to anyways.
1
100
