Vassilios Karakoidas

In 2023, we spent $3,934,099 on AWS + other hosting. In 2026, our hosting + support bill is down to ~$1m/year due to the cloud exit. Even including all the hardware buying, we will already have saved ~$4m by the end of this year. And going forward, it's ~$3m/yr in savings 🤑

Unlike Bob, I review all code generated by agents. Test coverage and similar metrics will give me confidence of functionality, but they offer me no confidence whatsoever that those agents have not introduced vulnerabilities, that they have not introduced dead code that will diminish understandability in the future, that they have missed factorizations that would have significant impact upon performance. Trust but verify. As an experienced developer, I know the smell of what is good and what is not. And no agent has either the experience or the context to know those things. If you want to be sloppy and fast then I suggest you proceed with Bob’s advice.

i am happy that anthropic is doing what they are doing with subscriptions the real cost of ai is really high and this subsidized life coming to an end more and more is a good thing.

Why your AI prompts fail and how to fix them ... @bkarak/why-your-ai-prompts-fail-and-how-to-fix-them-a66fd7bbe409" target="_blank" rel="nofollow noopener">medium.com/@bkarak/why-yo…

How long do you think the AI companies will subsidize its usage? Now, they get meta-tons of training data. When they have enough your typical 200$/mo program will jump to 2000$/mo.

I find it so wrong to just copy #Claude #code or use a copied version.

The prompt scorer that doesn't spy on your prompts. 7 merit dimensions, deterministic analysis, zero retention. promptivo.wizhut.tech #prompts #ai

@LiorOnAI All these are really useful, but it is productization of the models. Not real innovative. Brute force approach. Love them, cannot argue with that though.

Dario Amodei in 2025: "In 12 months, we may be in a world where AI is writing essentially all of the code."" Anthropic 2026: - Jan 2026: Claude Cowork launched. - Feb 2026: Opus 4.6 released. - Feb 2026: Sonnet 4.6 released. - Feb 2026: Cowork launched on PC - Feb 2026: PowerPoint integration - Feb 2026: Excel integrations added. - Feb 2026: Co-work plug-ins released. - Feb 2026: Claude Code security launched. - Feb 2026: Claude Code Remote Control - Feb 2026: Scheduled Task in Co- work - Feb 2026: Connector available in the free - Mar 2026: Claude memory is free - Mar 2026: Claude Marketplace launched - Mar 2026: Claude com ambassadors - Mar 2026: Code review for Claude code - Mar 2026: Claude skills for Excel & Slides - Mar 2026: charts & diagram in chat - Mar 2026: 1 million context window - Mar 2026: Dispatch for Claude Co-work - Mar 2026: Claude code Channels - Mar 2026: Co-work Projects - Mar 2026: Claude Computer use - Mar 2026: Auto mode in Claude code.

what the ... getfresh.dev #atlast

I am reminded of AlphaGo's move 37, an action that was completely unexpected yet sublime. Matthew is right to be impressed by his collaboration with Claude: in the hands of a expert in a domain, tools such as it are quite useful The operative words here are "collaborate" and "tool" and "domain expert". Read his report and you'll find that this was indeed a vibrant back and forth between an expert and a tool, a domain expert who could quickly discern right from wrong, an expert who could direct the tool not unlike a seasoned surgeon can feel the texture of an organ as she moves a scalpel over it. Here's the parallel to AlphaGo: in both cases, we have an AI that is unrestrained by history or tradition; in both cases AlphaGo and Claude have at their disposal a massive network that has been unexplored by human minds, and so in the right hands, a human can quickly follow paths that yield new mountaintops and placid new valleys. Are these things AGI or even a path to it? Are they capable or reasoning or understanding? No. Absolutely not. But - in the words of Steve Jobs - they are a bicycle for the mind.

What are going to build today? #ai

Neuralink is restoring speech to those who have lost the ability to speak

The internet is dying

@sama There are people in your team that still do. You can ask them. :)

I have so much gratitude to people who wrote extremely complex software character-by-character. It already feels difficult to remember how much effort it really took. Thank you for getting us to this point.

This is really how the AI era will go as it seems. We have to work on this problem.

It seems that google reduced the pro plan to a few (1-2) hours of coding. Just made the plan unusable for this task.

We are living this, but with broadband internet :) youtube.com/watch?v=-1F7va…

Many talented people over the past few years were declined an offer or even an interview @xAI. My apologies. @BarisAkis and I are going through the company interview history and reaching back out to promising candidates.