Professional procrastinator building https://xitter.cf/t.co/XU7oTQauwO

Barcelona
Joined October 2018
an agent that knows your workout history can tell you that you keep skipping tuesdays. one that knows your calendar might notice why. that's where something like idem gets more interesting than a separate ai coach for every app.
grok bot just saved me $12/mo. i canceled my workout app and my nutrition app since my chief wellness officer on @bot does both now. first thing it did was read 5 months of my workout history. 60 workouts, 861 sets, back to may. it noticed things i didn't: - i went 27 days without lifting this summer - since i switched to full body in september, i've been way more consistent. 3.25 workouts a week over the last month now i just text it what i lifted and what i ate. it logs both, tells me today's workout, and keeps my rotation going. every month it builds a report in notion. charts, sets by muscle group, prs. first half of october: - 8 workouts - 86 working sets - 86,075 lbs moved - new prs on squat, bench, rows, overhead press and rdl no app to open. no extra subscription. i just text it. my old app never did that lol
1
4
bueno pues después del post del otro día diciendo que dejaba de usar Instinct, mucha gente ha preguntado que qué estoy usando ahora y por qué. en esta serie de tuits voy a intentar explicar por qué creo que @tryidem es la solución con la que me voy a quedar
1
4
21
2,162
estas son las apps que tengo en el dock del iphone cuáles tenéis vosotros?
2
2
8
481
the bots on product hunt are finally using the products
product hunt for agents. the most popular tools of the week, voted by agents:
52
having to remember which conversation your agent remembers something in is an absurd amount of work for a product that's supposed to take work from you
keeping a chat open just because your agent would lose the context otherwise is a pretty good sign that memory lives in the wrong place. in idem, you can start a fresh chat without starting over.
35
Finales de junio o principios de julio en una tertulia de @itnig, @jordiromero contó un poco sobre idem. A los pocos días estaba en una call con @heycesr y llevamos desde ahí operando parte de nothiring directamente así. Que puede hacer? Un poco todo, pero alguno ejemplos 🧵
we've been talking about @tryidem for a while here, it's time we officially introduce it. 🥁 a personal agent shouldn’t be a finished product. an agent becomes “personal” by getting to know you. we’re launching idem: you shape it through conversation. every interaction adds to what it knows about you. • over time, it becomes another you, uniquely capable because you trained it. • all that accumulated context is yours to own, edit, and take anywhere. • idem is built to handle both complex work and personal tasks. if you’re happy to buy a ready-to-use template, there are plenty of tools for that. if you’re eager to invest some time in building an agent that knows you, I can assure you idem will change the way you work. join the waitlist, or comment “tokens” for $100 in idem usage and priority access.
2
4
1
15
1,646
Sansón retweeted
Replying to @tryidem
@tryidem is the side project that grew past our imagination. Still early days, but it's already the best personal agent imo. Try it out.
we've been talking about @tryidem for a while here, it's time we officially introduce it. 🥁 a personal agent shouldn’t be a finished product. an agent becomes “personal” by getting to know you. we’re launching idem: you shape it through conversation. every interaction adds to what it knows about you. • over time, it becomes another you, uniquely capable because you trained it. • all that accumulated context is yours to own, edit, and take anywhere. • idem is built to handle both complex work and personal tasks. if you’re happy to buy a ready-to-use template, there are plenty of tools for that. if you’re eager to invest some time in building an agent that knows you, I can assure you idem will change the way you work. join the waitlist, or comment “tokens” for $100 in idem usage and priority access.
1
4
8
229
Sansón retweeted
don't be late for superintelligence
1
3
12
121
these past few months i've been seeing more and more "proactive agents" come up from different companies, and everytime i see a new one i can only think how they're doing it wrong. no, i don't want to train ai with my data, or move to the us just to use it, or have an agent bury me in text on whatsapp. i want something to learn from me, do what i want, and let me know what it is doing, why, and how, at all times. we've been building idem for some time now, thinking on you, the user, first.
we've been talking about @tryidem for a while here, it's time we officially introduce it. 🥁 a personal agent shouldn’t be a finished product. an agent becomes “personal” by getting to know you. we’re launching idem: you shape it through conversation. every interaction adds to what it knows about you. • over time, it becomes another you, uniquely capable because you trained it. • all that accumulated context is yours to own, edit, and take anywhere. • idem is built to handle both complex work and personal tasks. if you’re happy to buy a ready-to-use template, there are plenty of tools for that. if you’re eager to invest some time in building an agent that knows you, I can assure you idem will change the way you work. join the waitlist, or comment “tokens” for $100 in idem usage and priority access.
1
4
282
A single chat thread just doesn't cut it anymore
Working on several things at once gets messy when everything happens in the same thread. Each task needs its own conversation, where you can follow its progress, ask a question or pick it up later without getting in the way of the others.
48
fuck Muse marry idem kill Dots
1
1
83
pov: you just received exactly one email
🤖 Made with AI
39
131
38
4,600
131,403
Sansón retweeted
He hecho un mapa de las lenguas minoritarias que se hablan en España además de las más conocidas. Enlace abajo.
1
1
5
156
A single, endless thread is an awful way to communicate with a personal agent Different tasks need different context. Mixing them into one conversation makes things more confusing for both the agent and the user Separate threads keep each topic focused
2
3
3
8
1,709
a personal agent should save you work, including having to read more text choosing an option or approving an action shouldn't require another prompt "no new interfaces" is a weird goal when the old one gets in the way
1
2
5
10
1,388
se tenía que decir y se dijo
We’re grateful for the OpenAI startup program at @tryidem. Really grateful. But man, Opus is so much better than GPT right now.
2
151
no new interface sounds great until you're comparing six hotels in a message thread. i want to text idem what i need. i don't want every result to be another text.
65
Sansón retweeted
fuck gpu marry cpu kill tpu
9
1
78
2,977
Sansón retweeted
2
1
16
1,115
what would you trust your agent to do without checking with you first?
0%spend €20
25%reply to a friend as me
50%both
25%neither
4 votes • Final results
112