Anthropic's strategy seems to go towards "AI as universal glue". They want to tie Claude into all the tools teams already live in (Jira, Confluence, Zapier, etc.). That's a smart move for enterprise adoption, but it also feels like they're compensating for a plateau in core model capabilities.
Both OpenAI and Google continue to push the frontier on reasoning, multimodality, and efficiency whereas Claude's recent releases have felt more iterative. I'd love to see Anthropic push into model research again.
I am sure they are already doing that. To think that an AI researcher is doing essentially api integration work is a bit silly. Multiple efforts can happen at the same time
They certainly have internal research efforts underway, but I'm talking about what’s actually been released to end users via the Claude app or API. Their latest public Sonnet release 3.7 (feb 2025) felt pretty incremental compared to Sonnet 3.5 (june 2024), especially when you compare them to OpenAI and Google released models. In terms of the models you can integrate today, Anthropic hasn’t quite kept pace on either reasoning performance or cost efficiency.
Anthropic is now too expensive to be acquired. Only Amazon could be a potential buyer, given that out of the 3 big cloud providers, it's the only one without their own model offering.
What they are doing is taking the model and elevating it to the level of attention controller in the conversation. It can run tools, spout inferences, etc. If they can control the UI in the CLIENT, they win. If they don't, they have to rely on someone putting their API integration into another project.
I’ve tried several AI code reviewers in the past few months and honestly, most of them fall short in the same ways - they lack proper context about the codebase and can’t remember discussions from previous PRs or design decisions.
IMO the tools that will win this space will be the ones that properly understand the codebase context and coding/architecture patterns.
Btw are you also planning to support customizable rules for code reviews?
Hey thanks! You re spot on, understanding the codebase context is key and maintaining history of the repo is perhaps the top problem to solve going forwards.
Also, there is context that’s not just code but documentation/guidelines etc, plenty of experimentation to be done there.
Customisable rules? YES - i ve got a lot of plans currently on this. I see two layers of customisation:
- allow the user to customise when and how often reviews happen, and how they re triggered (i.e every commit/every x mins etc, avoid running prs on specific files I.e package.json changes etc)
- pass in custom rules to the reviewer that work well for you(or team in the future). This will allow the tool to be more dynamic in nature as opposed to just LLM + your code.
Hope this answers your questions, thank you for the support!
I 'll disagree with most comments that it's mainly an organizational problem. Creating tooling for things like:
- managing different data sources
- versioning data
- monitoring how new data affects the model
- testing that certain SLAs are met before new features are deployed
- ability to rollback
- data & model quality monitoring
is technically challenging.
Obviously there are engineers that will quickly hack something together and will falsely think that
they have a good enough MLOps solution. I have been part of such teams.
Most companies are not Google, Facebook, or Uber. The large ones very often don't have the know-how to create
a robust technical solution around this process, and even if they do it can take them years and the smaller ones
lack both the resources and technical expertise.
I'm always looking for new ideas that can become successful business and when I saw the Uber Michelangelo
here on HN a few years ago, I was thinking that selling similar tooling to other companies, had great
potential. Seems that the right team to create that company was the one the built Michelangelo itself :)
Appreciate your feedback. Having shared for several years a house with other people (some of them friends, others people that I was meeting for the first time), we faced problems which were not related to communication only (that can easily be solved with a chat app and sticky notes).Here are a few examples:
- Forget/delay to pay the bills
- Couldn’t arrange to do the house cleaning in a time convenient for everyone due to different schedules. In the end half the people would skip that task.
- When I was sharing a house with people that I was meeting for the first time, there was a need to socialise and plan to do some stuff together. We found out that using an app for social planning was very helpful and we could also get some ideas on what to do.
- Often I would get back late from work, when the supermarket was closed, so I needed someone to do the shopping for me. To solve this problem I was sending screenshots of my shopping list to the house group on whatsapp and I would ask my flatmates whether they could buy me that stuff.
And these are just a few. From my experience when you are out of the house most of the day it is difficult to solve these problems only with a chat app and sticky notes. Often we would use a different app for each problem. So me and my partner decided to build RoomEase.
Before we start developing this app, we pitched it to potential users. So far we have feedback from around 100 potential users and ~90% was very positive (some of them are friends, so we didn’t expect negative feedback on those cases). This gave us enough confidence to move on.
Regarding your last question, you probably mean different target audience. The market we are targeting is the market of house tasks and services, which is a validated one.
Once we get our web app online, if we don’t see much growth on the number of users, we will reconsider the idea and who our core users are.
I understand that the demo is probably not very representative of what I am mentioning on the examples, but it is just an mvp. Again appreciate your feedback, it is good to know what people are thinking of this app.
Thanks for your question. That gives me the chance to clear thing up. It's going to be a long answer, but seems that my initial description caused some misunderstanding.
We are building a platform that targets people who are living in a shared property. So tenants will be our core users. Once signing up, a user will either create a new group or join an existing one (invitation only - users won’t be able to discover other groups). The group represents the house and inside that group only the tenants can interact, chat, do social planning, etc. It is a private place, where all the interaction between the users will remain private and they can’t be reached by people outside the group. That’s the core part of our app.
Now the second part is the marketplace of tasks and services, where people that have some skills or can offer some services and want to get paid for that, can list themselves. You can think of it as something similar to TaskRabbit.
Why are we building that? Because while interacting with their flatmates, people may notice that need some help on house duties or other tasks. For example you may want a professional cleaner to take care of the house cleaning, or you may realise that you don’t have time to do the shopping and want to find someone locally that can buy what’s on your shopping list. In that case, you go to the marketplace and ‘book' a person that will accomplish your task.
Finally the landlord will also not have any access to tenants’ activity. However we want to make it easy for flatmates not only to communicate with him, but also to raise any issues occurred in the house, such as damages, so that the landlord can take care (or even him can go the marketplace and search for a handyman that will do the fixes). We haven’t done any work on the landlord’s side, yet. These are just some ideas.
So that’s all! Please let me know if you have any further questions.
Both OpenAI and Google continue to push the frontier on reasoning, multimodality, and efficiency whereas Claude's recent releases have felt more iterative. I'd love to see Anthropic push into model research again.