chain agnostic▫️ai automation▫️growth lead▫️social media and community
Republic of the Philippines
Joined August 2022
- Tweets950
- Following72
- Followers93
- Likes2.2K
Ej | Web3 | AI | Crypto retweeted
Revolut: Europe's most valuable startup ($200B) just showed the world its AI strategy.
Every company will end up copying it.
At the start of this year, Revolut leaned on the frontier labs. Its 2,000 engineers used Claude Code and Codex, and some burned $1,000 to $10,000 a month each. The bills kept climbing, and the returns weren't there.
So Revolut switched its default to an open-source model called GLM. All it needs is GPUs, which it rents from Together AI.
The results, by Revolut's own count:
- The bill is about a tenth of what it paid OpenAI and Anthropic
- Engineers get unlimited tokens instead of spending caps
- Coding time is down 60% and lead time is down 70%
That's the chart below.
Revolut still uses Claude and GPT, but only as "the exception," for the tasks they do best. Everything else runs on open source.
So here's the question I keep coming back to: how many use cases can companies really justify sending to frontier models?
My answer is very few. Plenty of companies still route everything to OpenAI and Anthropic today. For how long?
I think open source ends up handling the vast majority of corporate AI work, and frontier models get a short list of the hardest jobs. That's a problem for labs whose revenue forecasts assume unlimited growth.
If you appreciate unpopular takes, follow me @m0xt_ for more.
ALT Revolut chart of engineering lead time and coding time, both dropping after switching to self-hosted open-source GLM
Ej | Web3 | AI | Crypto retweeted
BITCOIN JUST BROKE ABOVE $84,000 🚀
If this level holds and we close the week above $83K, this could be the start of the biggest bull run.
Ej | Web3 | AI | Crypto retweeted
I use Claude and Codex models pretty heavily for work.
When I build applications, I usually reach for OpenAI, Anthropic, or even Gemini APIs sometime for low balled tasks xd.
The problem is that managing multiple providers gets too big of a task alone -
>Different API keys.
>Different billing/recharges.
>Different SDKs.
>And if you're following Hexagonal Architecture for agents like me, another adapter for every provider.
I recently tried Distokens to solve this provider + cost + memory + build problem.
The architecture changed quite a bit:
> One adapter in my agent architecture instead of one per LLM provider.
> OpenAI Agents SDK everywhere, while Distokens handles the model providers underneath.
> Task-based provider selection, so I can use the right model for the job instead of defaulting everything to the most expensive one.
> Persistent agent state during development. I'm saving work summary where supported and agent memory in .md files ( using my cerebra - brain styled memory system underneath hehe ).
So when I switch providers, the next model doesn't have to start completely from scratch. It can pick up from the state I've persisted.
And for agentic workflows, it makes running REACT/OODA-style loops a lot more practical without burning through expensive models on every step.
Definitely worth it for me, do give it a try, sign up via my referral link and get $20 worth of credits for freeeeeeeee -
distokens.com/register?promo…
🤝 Paid partnership
Every AI provider eventually hits you with something: an outage, a rate limit, a pricing change, or a model that no longer fits your workload.
Building your entire AI stack around one provider makes that your problem too.
Distokens gives you one @OpenAI-compatible API for hundreds of AI models, so you can switch models, optimize for cost and keep building without being locked into a single provider.
One API. Hundreds of models. More control over your AI stack.
Build with Distokens.
Ej | Web3 | AI | Crypto retweeted
LATEST: 🤖 OpenAI CEO Sam Altman says AI compute is "probably already or at least rapidly becoming the most expensive infrastructure project in history."
(Source: David Senra)
Ej | Web3 | AI | Crypto retweeted
🔥 TODAY: Elon Musk's SpaceX has partnered with Nvidia to build a space-optimized Vera Rubin NVL72 system, set to launch to orbit in Q4 2027 with major scale-up planned by 2028.
Ej | Web3 | AI | Crypto retweeted
Tokenization will likely be the biggest financial upgrade of our lifetimes.
Ej | Web3 | AI | Crypto retweeted
🚨 JUST IN: Meta is set to launch its consumer AI agent platform "Hatch" within the coming weeks, ahead of its new "Watermelon" model in October, per The Information.
Ej | Web3 | AI | Crypto retweeted
🛡️ Protect your account from phishing and link spoofing:
🟦 Set your Anti-Phishing Code in Settings
🟦 Turn on 2FA, not SMS
🟦 Bookmark the real site and use the bookmark
🟦 Keep your app and device updated
Ej | Web3 | AI | Crypto retweeted
"Which AI model is cheapest?"
Useful question.
But usually not the best one.
A $0.01 request that requires three retries can be worse than a $0.02 request that works immediately.
A cheaper model is not cheaper if it increases:
→ retries
→ latency
→ context usage
→ agent loops
→ engineering overhead
What matters is the economics of the whole workload.
That's part of why we're building Distokens around multi-model access rather than a single-model bet.
One @OpenAI-compatible API. Hundreds of models. More freedom to optimize around the workload.
#Distokens #AIInfrastructure #AIEconomics #AIEngineering
Ej | Web3 | AI | Crypto retweeted
We’re giving scientists, mathematicians, and engineers free access to our frontier models—starting with 10,000 researchers and expanding to 100,000 through 2027.
ChatGPT for Academic Researchers is built to accelerate discovery across disciplines.
Ej | Web3 | AI | Crypto retweeted
You're probably paying more than you should for AI APIs.
Distokens helps you:
• Auto-route to the best AI model
• Save up to 30% on API costs
• Keep the same OpenAI SDK with zero code changes
• Estimate your savings using the Distokens Cost Calculator
Retweet this post, follow @ai_distokens, and DM us to receive up to $50 worth of free AI token credits.
#Distokens
Ej | Web3 | AI | Crypto retweeted
Most AI apps don't actually need the same model for every request.
Paying premium pricing for every API call feels outdated.
Distokens caught my attention because it automatically routes requests to the most suitable AI model using the same OpenAI-compatible SDK, so you don't have to rebuild your workflow while potentially lowering costs.
you can claim up to $50 in AI token credits to try it yourself.
✅Retweet the main post
✅ Follow at @ai_distokens
You're probably paying more than you should for AI APIs.
Distokens helps you:
• Auto-route to the best AI model
• Save up to 30% on API costs
• Keep the same OpenAI SDK with zero code changes
• Estimate your savings using the Distokens Cost Calculator
Retweet this post, follow @ai_distokens, and DM us to receive up to $50 worth of free AI token credits.
#Distokens
Ej | Web3 | AI | Crypto retweeted
one thing more AI builders should optimize isn’t prompts.
it’s infrastructure.
the model matters, but how your requests are routed can have a bigger impact on cost than most people expect.
this is worth checking out if you’re building AI products.
retweet the original post, follow @ai_distokens, and DM them to claim up to $50 in free AI token credits.
🤝 Paid partnership
You're probably paying more than you should for AI APIs.
Distokens helps you:
• Auto-route to the best AI model
• Save up to 30% on API costs
• Keep the same OpenAI SDK with zero code changes
• Estimate your savings using the Distokens Cost Calculator
Retweet this post, follow @ai_distokens, and DM us to receive up to $50 worth of free AI token credits.
#Distokens