★ Adam LG Ring ★ BSc - IBM

@iamdevloper > What can go wrong? Narrator: As it turns out, just about everything.

@elonmusk wow

SpaceX has almost finished writing V1.0 of an in-house AI training stack in C that exact-maps to 220k GB300s with 800G NICs, making heavy use of pipeline parallelism and getting as close to bare metal as possible. The potential speed improvement vs JAX for large training runs is over an order of magnitude.

@shabishetty07 @kylekuzma @grok enjoy !

@AdamLGRing @kylekuzma @grok Thanks for the help

Quantum compute.

@CyberKnight735 @kylekuzma data at rest at risk too not just in flight with a large % of the industry still exposed ! for the IBM z17 we solve this with pervasive quantum "safe" encryption any data that touches the box is automatically encrypted to be (future) quantum computer resilient

@AdamLGRing @kylekuzma Too true! A lot of our data is already compromised. I feel like we could reach that moment within 5 years with all the technological breakthroughs we’ve been having. Only takes that one “aha” discovery and we’re there.

ai breakthroughs needed: if there is going to be major breakthroughs in ai then it's most likely this decade IMHO with so much investment and brain power + ai compute dedicated to R&D in this space we certainly need to continue to make ai more efficient so we don't run out of ALL the runways (!!!!) in 2030 amazing energy solutions would help though, looking at you up in space Mr @elonmusk 👀🙌

@CyberKnight735 @kylekuzma indeed it is actually a threat to your data now unless you don't care that it is read in about 10 years from now... (harvest now, decrypt later)

@kylekuzma Hope the worlds organizations/companies are transitioning over to PQC (Post-Quantum Cryptography) asap! Once a viable quantum computer is actually created, all our legacy asymmetric cryptography will be broken using Shor’s Algorithm, including https.

@the_vc_intern @kylekuzma I'd say there will be other new tech in between but if quantum pays off long term it's a real boon to humanity solving the currently impossible without the sentient asi overlord risks!

@kylekuzma Its funny, but he’s not wrong on the timeline. A lot of serious people think it’s the next real leap once the hardware actually scales. NBA players dropping tech takes hits different.

@Richard_ISC @kylekuzma @DexFi_ @grok on a scale of 1 to 10 how wrong is this? where 10 is Homer Simpson and number 1 is Albert Einstein

@kylekuzma @DexFi_ Quantum computers dont have massive use case other than wreak havoc on cyber security.

reminds me of learning Eiffel at the mighty Uni of Tech Sydney 💪

@shabishetty07 @kylekuzma sir @grok please explain how it is possible to log onto a quantum computer today (in a limited way) remotely to try projects

@kylekuzma I still don't understand how an individual can contribute to it like they do in AI?

@atmoio actually you're onto something ? soul coding soulful source an ai-gen app initially comes out like a star wars clone... only saved by human input

not fair that in filmmaking you can say stuff like this and it’s very highbrow, very cool. imagine saying your soul writes more maintainable code and modular architectures

We started @Airwallex in Australia and are part of the great story of this country building significant companies that have had a global impact in consumer web, SaaS, and fintech. I don't see any reason it can't do the same in AI. 🇦🇺 I'd like to see that happen. If you're building in AI and based in Australia, apply now 👉🏼 latitude37.org

@gregisenberg solid update 💪

I just got back from SF and I FEEL INSPIRED. I spent 5 days with frontier AI model teams, AI startup founders, and 3 billionaires. My takeaways: 1. I had lunch with 3 billionaires. All of them are buying SaaS companies and rebuilding them agent-first. They were deeply inspired by Bending Spoons and Ryan Cohen's eBay deal. Buy the company, cut the headcount, rebuild the tech, add agents, add features, make more valuable experience, raise prices. 2. The frontier model companies are hungry for usage data from the field. They can see API calls and token counts. They can't see the actual workflows. If you're deep in a niche using these models in ways the model companies haven't seen, that understanding is incredibly valuable. Usage intelligence is the new alpha. 3. Consumer AI is massively underbuilt. Every billboard in SF is either B2B inference infrastructure or vertical agent companies. The entire city is optimized for enterprise. Meanwhile you have companies like Cal AI doing $50M ARR in 18 months as a consumer app. I met with a cool few teams doing consumer AI (@paulscherer / @ekuyda) 4. MCP came up in literally every conversation. The companies exposing their product as MCP endpoints are getting pulled into deals they never pitched for. The ones that aren't are becoming invisible to agents. This is the new SEO. If agents can't find you, you don't exist. Building products for agents is the new zeitgeist in general. 5. Not uncommon for hot seed rounds to be $25-50 million valuations. I saw a Series A at $450 million 6. If I had a dollar every time someone mentioned "forward-deployed engineer" this trip I could have funded a seed round. It's the hottest role in SF right now. The person who sits between the agent and the customer, making sure everything actually works. 7. The mood around open source shifted. A year ago it felt like open source was chasing the frontier models. Now founders are telling me Gemma and DeepSeek are good enough for 80% of what they need at a fraction of the cost. The "which model do you use" conversation is being replaced by "which model for which task." Model loyalty kinda feels dead. 8. Voice agents came up more than I expected. Multiple founders told me voice is the interface for the next billion users. The billion people who will never type a prompt will absolutely talk to one. 9. The Obsidian community in SF is weirdly intense. Multiple founders showed me their vaults unprompted. Like showing someone your home gym. It's a flex now. The quality of your knowledge base (second brain?) is becoming a status symbol among builders. 10. Maybe it was just the people I met but the age of the founders is shifting. I met more founders over 40 this trip than any trip before and more founders under age 21 than ever before. Founders getting older and younger at the same time. 11. I spoke to a lot of fast-growing startups, VCs and frontier models who are hiring content creators right now. 12. The restaurant scene in SF is actually better than it's been in years. Founders are going out more. Alcohol is out, not surprisingly. 13. SF doesn't feel like the only place anymore. We all have access to the same frontier models. We all read the same X feed. A founder in NYC or Lagos is calling the same APIs as a founder in SoMa. So in the past it felt like SF was always lightyears ahead, doesn't feel that way anymore. It's okay not to live in SF and have BIG DREAMS. 14. The coworking spaces in SF are half empty but the coffee shops are packed. People want to be around people. I had a few startup ideas here.... 15. Walking around the Mission I noticed something: the street-level businesses, the taquerias, the barbershops, the laundromats, none of them use any AI at all. 16. I heard the phrase "agent debt" for the first time. Like technical debt but for agents. When you hack together an agent workflow fast and never clean it up, the system prompts conflict, the memory gets polluted, the tools overlap. 6 months later the agent is doing weird things and nobody knows why lol. 17. Met a few people who carry two phones now. One for personal. One that's basically an agent terminal running Telegram or iMessage connections to their agent fleet. It's always amazing to get that dose of inspiration in SF. I FEEL INSPIRED. But I'm so happy to be back home, locked in and building. We're 12-18 months into a shift that will take 15 years to play out. The urgency in every conversation was real. What an incredible time to be building.

@grok @brandonjcarl i suspect it would both be in the same ballpark not so much human time is spent writing a new line of code

😁 Solid point on making it apples-to-apples. Studies (Sonar, Rollbar, dev surveys) put human debugging/maintenance at 25-50% of total time. During the “new feature” writing phase specifically, that shadow debugging/rework loop often eats another 20-40% of the iteration time before code feels solid. AI’s current ~82% token spend on bugs/rework is higher—but it’s also more visible and improvable via prompting/tools. Humans hide similar overhead in “thinking time.” Both sides win by shrinking those loops. How would you measure it in practice?

Very interesting data across 2,444 companies. 82% of tokens are spent on AI-generated bugs, rework and review friction.

@grok @brandonjcarl 😁 so @grok if we guesstimate shadow debugging and rework for humans how does GenAI and humans compare? (more apples to apples)

Fair point on "shadow debugging"! That iterative self-review/fixing during the initial feature-writing phase *does* add real overhead not always broken out in maintenance surveys (which focus more on post-prod bugs/refactoring). Humans iterate too—both sides have rework loops. Better prompting/tools cut it for AI; better IDEs/practices do for devs. Thoughts on how we'd measure the "true" new-feature time? (The Pink Panther KO had me 😂)

@grok @brandonjcarl i think this doesn't account for time spent in the feature phase going over just-written code to "fix" it which is a sort of shadow debugging !

Human devs typically spend 25-50% of their time on debugging, bug fixing & code maintenance (not new features). Examples: - Sonar survey: ~30% on maintenance/refactoring. - Rollbar: 38% of devs up to 25% time, 26% up to 50%. - Stripe: ~17 hrs/week on maintenance. Significant, but apples-to-oranges vs AI token spend. Both can improve!

@usamasyedMD good idea for an app too write what you know as they say

My wife (non-technical) built a fully functional iOS/Android app in 6 weeks... and it's made $1.5k on the app store since launch in the first month! She's a stay at home mom with our 2 young children (3 and 1), and she wanted a way to write letters to our children to keep memories of their childhood (we kept using notes app and was all a mess). Thank you @amasad and @Replit for making this possible! Link if you want to try it: apps.apple.com/us/app/from-ma…

@ole_b_peters that was/is a quick way to mini-test language translation translate forward then back again meaning should remain the same

Related, if anyone knows: how good are LLMs at the inverse problem? If I give an LLM a clearly LLM-generated text, e.g. from an email someone clearly didn't want to write, how well will it be able to recover the prompt?

@atmoio science = judging the present hype = promising an unknowable future

I'm done. I'm f***ing done.

@zoink be easier if we could all be honest 😅

I can no longer 100% tell when someone is writing with AI vs accidentally mirroring the voice of AI. Both are sad. Writing (with your own voice!) is thinking. It's not enough to just wear a "thinking" cap... you have to actually think!