ashwini asokan

Singapore’s Foreign Minister, Dr Balakrishnan casually explaining how he built his own AI agent (a 2nd brain for diplomacy) using Claude & WhatsApp integration etc. on a Raspberry Pi “You cannot govern a technology you have only been briefed on.” 🇸🇬

@beingSathya 🤯🤯🤯

@LadyAshBorg Crazy quarter they got. 32% up with a huge $1 billion in revenue. I guess the more ai demand, its better for them

Everyone: Software is dead Datadog: Hold my beer!! What even is that stock price!?! 😲🤯

Refreshing read. The only irony here is that the system itself is designed to produce bad speed with something new coming up everyday that requires the same pool of talent and what we're getting is this below - everyone moving to the next shiny thing in 12-18 months. But as all cycles go, this is a peak cycle and we'll see a return to other parts of the hype cycle when AI actually starts getting productionized beyond the handful of use cases we see around today. "Bad speed is impatience that pretends to be ambition. It optimizes for novelty over depth and doesn’t let you sit long enough inside the unglamorous stretch where the real edge is formed."

OpenAI releasing their update on GPT Realtime stack says a few things about where the broader voice agent market is headed - Voice AI adoption is moving from early adopters into mainstream product behavior - “Voice as a feature” is coming with even more force to an app near you - “Voice as an operating layer” is going to see a ton of startups both competing with & building on top of Wispr, ElevenLabs etc shifting from just the voice quality arc to latency, workflow integration, personalization, realtime context and more I think we're at escape velocity with voice. Feels a lot like the early App Store era, but for conversational interfaces My tinies are grown up now but I'm certain we're going to be seeing a ton of these apps at home with the kids - storytelling and so much more without having to look at a screen!!

POV : USING CLAUDE OPUS 4.7 TO JUST RENAME A VARIABLE

How have we not monetized TN meme nation yet?!? What a force of life!

@prasanna_says Showed your non tamilness 😬😂

@LadyAshBorg Damnit showed my non gen-z ness

Why Vijay Anna should break his poll promise!? Six free cylinders sounds pro-poor => It is actually a ₹10,000 cr/year LPG trap! Give every TN household induction instead => ₹3,700 cr once! Cooler kitchens - cool women! No flame - safe from kitchen fires Less import dependence forever! Annapoorani Super-6 → Safe-6! @TVKVijayHQ

30 mins into the claude code keynote and every speaker so far has been a woman. just saying 🫶🏻 @asvora @angjiang @katelyn_lesse @_catwu Dianne Penn @claudeai

@vinaykesari @jaivardhan88 😂😂😂

The house always wins.

@LadyAshBorg was just watching the prediction marlets episode by John Oliver and this tweet pops up. Damn!

Prediction as a feature. Just came back home to my son placing bets on 2 players he's watching playing inside the game he's playing, to get some gems. Explained his predictions based on data he observed with each player. Slowly training our kids to be ready for the prediction markets I see!! (Noob gamer here so highly likely this has been around and I'm just noticing)

@sdhilip Its nuts. Absolutely awesome and nuts. The movie script writing part also I think tells us how prevalent this has gotten outside of our very loud tech industry bubble

@LadyAshBorg I was travelling yesterday from Bangkok to Kula Labour and the flight had WiFi. I saw at least 6 people using AI, a mix of Claude and OpenAI, to do their PPT, writing and other stuff.

Seemed like there was just 2 of us at this cafe where Im brunching by myself, doing some deep research on stocks with claude. Sure enough, guy at the table across is on a call collaborating w/someone writing a movie script using …. Claude 😂 there are now 4 people at this cafe

@thestalinism 😂😂😂 too much drama … movie claude has been giving major vibes and distracting …

@LadyAshBorg Did you mange to pick the stock? 😂

@noise_segar 😂😂

@LadyAshBorg 4 people means it should be renamed as "Quade"

Been discussing AI budgets at fast scaling startups both B2B and B2C for a couple of months now. And it seems like we're almost back at the SaaS budgets era of 2021 after the crazy purge and purse tightening in the last 2 years. AI marketing, Claude & vibecoding equivalents and 'experimental' buys. 3 main areas really for the most part now. But the SMB/mid market tool spends are back in its full form, and likely beating the SaaS era highs.

Love this. Seems like domain moats are underestimated right now. And we're overestimating AI agents absorbing domain knowledge centric workflows (often tribal know-how/unstructured flows), in the near term at least. How we actually solve / solution for today's sub-optimal workflows (with an understanding of domain work) is another example of strategic illegibility, something that companies can own by virtue of knowing what might seem like stupid inefficiencies to AI but is knowledge work that needs to be baked into how the product is built. So product design of agentic work. In other words, perhaps some inefficiencies are there for a reason and agentic design can keep some of that illegible. Regulatory, finance, compliance workflows where policy and rules determine how things can be, come to mind here.

@localteaparty 💯

Moral of the Story: Succession Planning