
Product - Startups - Culture | B2B AI SaaS GraphRAG FinTech 'score-takes-care-of-itself' guy | 1x founder Prv 1st Product hire @chargebee, learning to learn.
Joined April 2014
- Tweets2.1K
- Following3.6K
- Followers480
- Likes3.9K
Replying to @claudeai
@claudeai design is deprecating
We got to migrate to Claude Chats (as artifacts) :(
But chat does not invoke 'design' mode until you ask explicitly.
Keeps hallucinating even though necessary context is present (design.md, components.md)
#agenticai #claudedesign
2
49
Rooting for @noondesign @paper
to take this opportunity,
double down on marketing
Claude design within chats is 👎🏼
so you build the true #ux #ui #agenticdesign workflows and re-take over the world 🙏🏼
10
"build with AI"
"be an ai builder.."
so I did.
now I am stuck in ci/cd
reviewing PRs and merge conflicts
writing millions of LOC is easy, pushing to PROD isn't
#aiengineer #vibecoding #cicd #devops #solobuilder #productbuilder #qa #agenticai #agenticcoding
3
2
74
I agree these are common Tech Lead / Engg Manager related areas and very normal to address.
For someone who cannot write production code, this is starting to become a concern.
1
5
incredible share @vineerpasam
@Designer_Bri take a bow 🙇👏🏼
29
Wow, learning a new abbr today - "B2A" 👏
On a side note: Will this change reflect in VC market maps? just curious 🤔
#agents #b2a #b2b #b2c
@openanysi
1
2
50
Did you audit your claude md/skill files recently?
You wrote the claude.md, agents.md, context.md.
Added the skills.
But did you check whether the instructions are followed as expected?
#aiengineer #agenticai #hallucination #skills #claude #codex #cursor #llm #claudemd
3
2
64
Here are my audit findings:
* Rules are not always followed
* Skill file shd have lesser instructions
* frequently test the instructions/ guidelines, they are not enforced always
* New model launch = retest your prompts and rules etc, rules get missed or not reqd at all
1
12
Very soon these digital avatars are going to go away
we will ask for human version of everything
agent/ai version for free
human version is paid
like currently a (human) relationship manager is behind a paywall
the fatigue kicks in -> VCs pouring in $$
#digitalavatar #agenticai
1
1
40
Prashant retweeted
Repeat after me:
My job as a leader is not to make people successful.
My job is to create the conditions that give them the highest possible chance of success.
Everything else is up to them.
Prashant retweeted
Underrated life advice: Spend less time consuming other people’s lives and more time building your own. Pick up a hobby that makes you forget your phone exists. A full life is one of the best cures for an envious mind.
Prashant retweeted
AI is bad for tech outside of SF/NYC imo
Being a non SV/NYC "tech hub" is just shorthand for "we have good engineers". Toronto is an example.
GTM > Engineering is not new, but now with AI coding - that gap will widen further. Cities w/o GTM and scaling culture will fall behind
🧨 A rather hot take - PMs expected to be AI builders may not be the best use of time, always.
I understand PMs shd build & learn how genAI products work.
However, time spent building is not going to yield high.
Why?
#prodmgmt #productmanagement #aipm #AIEngineering src: google
1
67
Effective PMs build AI around their day to day ops to improve their craft -
> @OpenAI ChatGPT research agents
> prototyping with @claudeai Design
> process feedback from @Gong_io + @sentry MCP
> analyse metrics via @posthog /@Amplitude_HQ
> data analysis with @hexanalytics
1
1
27











