Co-founder @with_ako: AI teammate who gets work done over Slack/Whatsapp. Try free - withako.com. Ex-@deliveroo.

Bangalore
Pinned Tweet
Today, we're launching Ako: our AI teammate that completes work for you in Slack and WhatsApp. Ako works across your team, connects to all your tools and gets work done. Try: withako.com You've heard this pitch before, so I'll cover the principles Ako is built on:
19
9
46
11,043
I don't think chat interfaces are great for learning a topic. To make things stick, you need to see what you've learnt, and what's next.
126
The best view of where the world is going: Download codex, step away from your computer, give it any task, watch it complete it in like 10 mins Now imagine everyone has this, it’s everywhere and 100x fast
1
3
276
Start a Github repo for every project instead of the native "project" inside every AI chatbot. Automatically makes everything portable
1
197
If you are business owner, I'd be obsessing about how you can get more business from personal agents like Muse, Dots, Grok, Instinct
138
To summarise current AI adoption in one sentence: 1) Single interface (e.g. WhatsApp) makes it ridiculously simple to use 2) Users are so used to UIs that they don't know what to do with it
1
147
I'm literally building an AI employee and I don't have time to test all of them. Consider the average business owner who is trying to figure this shit out
1
1
1
280
Hey @Google why do I need to sign in to my Google Workspace like every 1 hour? This is supremely frustrating
1
1
394
If you get your agents to do stuff with browser use, you will see why cyber threats are going to 10x with AI within seconds.
105
Retention is the biggest problem with always on, proactive agents. Outside of power users, people don't know what to do with them. Yes, your agent can be proactive and should be. But be too proactive and I will block you. Opportunity is wide open.
1
2
164
Look at the average person around you. They are probably watching reels or YouTube shorts. And then ask: what will they want out of a personal agent?
1
145
Krishna retweeted
Watch this acceleration by Vithya Ramraj to overtake the Chinese athlete in the final leg! - Helped India win its first Athletics Gold Medal at Asian Games 2026! 🥇 🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥
Sony Sports Network
804
10,000
46,943
1,442,515
Damn, they shipped a Jev competitor Expected but didn’t think it’d be this quick
Decisions API, for lightning fast constrained decision making powered by Luna. Supports visual inputs, and tuned to be able to make decisions in less than a few hundreds of milliseconds end to end.
3
211
Krishna retweeted
So many people in tech obsess about high income, metropolitan issues: dinner reservations comparison shopping luxury items travel booking These problems are just not that big for most consumers.
52
40
842
52,006
Krishna retweeted
Meta ads is a good example of a site that is made 100x better by MCP. The UI is an absolute abomination. Letting an agent do everything is so much better for your sanity.
105
9
795
48,304
Anthropic pls ship a model that is as cheap as Luna Sonnet, in our testing, is way faster than Luna (surprising?) Give me that speed for a lower price. I’ll take less capability
1
125
AI make you hyper optimise without actually retaining information. You can go down 100 rabbit holes, ask loads of questions and still not retain anything. You need to pay attention and listen, which the UX makes difficult because it's so easy to ask the next question.
3
186
This is a real interaction from a inside a Codex project. Made me giggle
1
130
So hard to learn something new as you get older.
learning a new language, after something like 30 years it’s so hard
137
An equally important consideration is who you are building for and what your north star is. If it's productivity and time saving, it's going to appeal to a small % of the population. For everyone else, they care about different things: making more money, jobs, fun
“Instinct for India” ≠ a conversational WhatsApp agent I keep seeing the two confused in pitches. WhatsApp is a sensible interface. What interests me about a personal agent is that the state of my life survives the conversation: what I’m trying to do, what I’ve authorised, what’s changed & what’s unfinished If an important meeting gets cancelled, does it remember why it mattered, find another slot & follow through within the permissions I gave it? Or do I have to brief it all over again? The same test applies in other categories like personal finance/quick-commerce/services. Research & price alerts can be useful. The harder product remembers why I bought, what would change my mind & connects new information to that context. It follows through on agreed next steps, asks for approval where needed & checks what actually happened That depth matters commercially. Otherwise, you’re likely squeezed b/w the broker that already has my portfolio/card/spending history & the horizontal agent competing to understand the rest of my life
146
You need to be proactive because the general population doesn’t know what to do & needs prompting. Do it too much or badly, and it feels like push notifications 2.0
And now it just feels like Instinct and Muse are stalking me with their endless proactive notifications Thirsty AI assistants
158