R.

2.9K posts

R. banner
R.

R.

@RichDoesTech

Christian | 1x Husband | 5x Dad | Building private AI tools that work offline @TryYaps (Sign up to the https://t.co/4B2vBqXwEl waitlist 👋)

London Sumali Ekim 2020
424 Sinusundan3.2K Mga Tagasunod
Naka-pin na Tweet
R.
R.@RichDoesTech·
# Day 1 Recently I've been focused on building a privacy-first and UX-friendly AI tool that handles a full suite of AI voice related tools (e.g. like, dictation, screen reading, and so much more) and slowly expanding into other domains. The hope is to address privacy, tool fatigue, latency, and several other things. But behind that the product, I'm challenging some assumptions that undergird the very root of how pre-LLM startups were built. I'll try and write more extensively on this. But for now, you enjoy the UI from my lastest Claude Code refactor 😅 (P.S. Shoutout to @DannPetty on reframe)
R. tweet media
English
5
2
28
2K
R.
R.@RichDoesTech·
Big step forward! Shipping @tryYaps to testers today on Android: - Low latency, offline dictation, with a lot less permission requirements than competitors - Seamless text-to-speech (optional) - Note taking, checklists, and Kanban boards integrated Hit me up for access
English
3
1
8
63
R.
R.@RichDoesTech·
@hrprtsingh9 "stubbornness is the only skill that actually matters" 🎯
English
0
0
1
10
Harpreet Singh
Harpreet Singh@hrprtsingh9·
Sold my first e-commerce business at 21. Made less than you'd think. Everyone congratulated me. Nobody asked what I actually walked away with after taxes, legal fees, and the earn-out I never collected. The number on the term sheet is not your number. The wire that hits your account 4 months later is your number. And by then you've already mentally spent it three times. Took that money and started AutoFlipz. India's auto aftermarket is a $15B mess. Unorganized. No standards. Customers getting ripped off at every garage. Built an AI-powered marketplace to fix it. That business taught me something the exit never did: building something you care about makes you 10x more stubborn. And stubbornness is the only skill that actually matters. XShore came after. $2.2M ARR. Different industry, different math, same stubbornness. Your first exit isn't the win. It's the tuition.
English
2
0
4
47
R.
R.@RichDoesTech·
@Fried_rice So now we can try and close that 20% gap between Claude Code and Cursor's harness open source 😅
English
0
0
45
8.7K
R.
R.@RichDoesTech·
@theo Lol, just confirmed it, that's interesting 👀
English
0
0
1
1.1K
Niels
Niels@niels_youtube·
If you're doing YouTube, start using Reddit for your research 👇🏼
English
1
3
48
2.2K
R.
R.@RichDoesTech·
@loganbartlett @tbpn Fair dues, I'd just expect someone with high enough agency to pick up most of that if not all prior. I was an actuary and so some overlap, but figuring the gaps out weren't crazy. I eventually just picked up Venture Deals and read it. I'm sure others would have too.
English
0
0
1
26
logan bartlett
logan bartlett@loganbartlett·
@RichDoesTech @tbpn good question. It was table stakes to have it for what the job required day one. Time to value mostly.
English
1
0
2
29
TBPN
TBPN@tbpn·
Redpoint's @loganbartlett says AI has completely changed hiring—favoring people with unique backgrounds: "Agency might be the only thing that matters." "That's the thing that we are trying to figure out—where do you find pockets of people who still want to do the job talent-wise, or have the capability to do the job, but also have agency?"
English
6
9
112
21.8K
R.
R.@RichDoesTech·
@ggerganov Beautifully written, completely in agreement. Surprising how fast the OS space has become viable on all sorts of devices. 2026 is gonna be a fun one. 🫡
English
0
0
0
138
R. nag-retweet
Georgi Gerganov
Georgi Gerganov@ggerganov·
llama.cpp at 100k stars now that 90% of the code worldwide is being written by AI agents, I predict that within 3-6 months, 90% of all AI agents will be running locally with llama.cpp 😄 Jokes aside, I am going to use this small milestone as an opportunity to reflect a bit on the project and the state of AI from the perspective of local applications. There is a lot to say and discuss and yet it feels less and less important to try to make a point. Opinions about viability of local LLMs are strongly polarized, details are overlooked, the scientific approach is lacking. Arguments are predominantly based on vibes and hype waves. One thing is clear though - local LLMs are used more and more. I expect this trend to continue and likely 2026 will end up being one of the most important years for the local AI movement. I admit that I didn't expect the agentic era to come so quickly to the local LLM space. One year ago, the available models were too computationally expensive for doing long-context tasks. There wasn't an obvious path towards meaningful agentic applications. The memory and compute requirements were huge. Last summer, with the release of gpt-oss, things started to change. It was the first time we saw a glimpse of tool calling that actually works well within the resource constraints of our daily devices. Later in the year, even better models were released and by now, useful local agentic workflows are a reality. Comparing local vs hosted capabilities at a given moment of time is pointless. To try put things into perspective: - We don't need frontier intelligence to automate searches and sending emails - We don't need trillion parameter models to be able to summarize articles or technical documents - We don't need massive GPU data centers to control our home appliances or turn the lights off in the garage I believe that there is a certain level of intelligence we as humans can comprehend and meaningfully utilize to improve our working process. Beyond that level, access to more intelligence becomes unnecessary at best and counterproductive at worst. I also believe that that level of useful artificial intelligence is completely within reach locally and it has always been just a matter of implementing the right software stack to bring it to the end user. With llama.cpp, I am confident that we continue to be on the right track of building that software stack! The llama.cpp project is going stronger than ever. With more than 1500 contributors, the project keeps growing steadily. From technical point of view, I think that llama.cpp + ggml is the only solution that actually makes sense. That is, the software stack must run efficiently on every possible device, hardware and operating system. The technology is too important to be vendor-locked. It has to be developed in the open, by the community, together with the independent hardware vendors. This is the only right way to build something that will truly make a difference in the long run. I won't try to convince you about what is currently and will be possible with local AI. We will just continue to build as usual. I am confident that after the smoke clears and we look objectively at what we have built together, the benefits will be obvious to everyone. Big shoutout to all llama.cpp maintainers. I feel extremely lucky to be able to work together with so many talented contributors. Every day I learn something new and I feel there is so much more cool stuff that we are going to build. Also, I am really thankful that the project continues to have reliable partners to support it! Cheers!
Georgi Gerganov tweet mediaGeorgi Gerganov tweet media
English
134
260
1.9K
146.5K
R.
R.@RichDoesTech·
@iamelijahkhan Never thought about it, assumed the US. Then when I saw the Cockney feature I started second guessing. The video confirmed it 😅
English
1
0
1
6
E L I J A H
E L I J A H@iamelijahkhan·
Day 14 of Accentify → Content creation pipeline is complete ✅ My OpenClaw, is managing a Notion pipeline of content, scheduling to Postiz and tracking analytics to maximise conversions + Released new accent course - Cockney. The raw, authentic sound of the East End of London
E L I J A H@iamelijahkhan

Day 13 of Accentify → $1200 MRR + 200 paying users! 🎉 We crossed $1000 MRR 9 days ago and now we're well on the way to $2000 MRR > Organic social media workflow w/ OpenClaw is complete (hopefully 🤞) > Unlimited UGC hook videos?! P.s count how many times I say "cooked" 💀

English
4
4
11
1.6K
R.
R.@RichDoesTech·
First thoughts on @omma_ai, really beautiful! One-shotted a functional animated background generator that I can use for screenshots, videos, etc. This is promising, especially considering I gave it no real visual direction.
English
3
2
16
784
R.
R.@RichDoesTech·
Big week ahead, bigger month ahead!
English
2
0
4
63
R.
R.@RichDoesTech·
When some bob, others will weave. Some will compete on evals, others on price. Some will fight to be the best, others the most affordable. Compute is almost certainly way more expensive than the current charges, agreed. But there's a misconception that everyone is willing to pay full price. I don't think that's true - and I do think companies will invest more in open source models or cheaper proprietary solutions over being at the mercy of ever-increasing prices. Also, the rising tide lifts all the boats. Models like Kimi facilitate the release of a Composer 2. This is a model that runs at like 1/20th the price of other frontier models, gets trained on real time data with a 5 hour post training cycle, and for coding (based on the evals) is likely "good enough" for most tasks. These divides between the various companies/labs will only continue to grow imo, leading to stronger divides in their respective ICPs.
Andrew Curran@AndrewCurran_

Three weeks ago there were rumors that one of the labs had completed its largest ever successful training run, and that the model that emerged from it performed far above both internal expectations and what people assumed the scaling laws would predict. At the time these were only rumors, and no lab was attached to them. But in light of what we now know about Mythos, they look more credible, and the lab was probably Anthropic. Around the same time there were also rumors that one of the frontier labs had made an architectural breakthrough. If you are in enough group chats, you hear claims like this constantly, and most turn out to be nothing. But if Anthropic found that training above a certain scale, or in a certain way at that scale, produces capabilities that sit far above the prior trendline, then that is an architectural breakthrough. I think the leaked blog post was real, but still a draft. Mythos and Capybara were both candidate names for the new tier, though Mythos may now have enough mindshare that they end up keeping it. The specific rumor in early March was that the run produced a model roughly twice as performant as expected. That remains unconfirmed. What is confirmed is that Anthropic told Fortune the new model is a 'step change,' a sudden 2x would certainly fit the definition. We will find out in April how much of this is true. My own view is that the broad shape of this is correct even if some of the numbers are wrong. And if it is substantially accurate, then it also casts OpenAI's recent restructuring in a new light. If very large training runs are about to become essential to staying in the game, then a lot of their recent decisions, like dropping Sora, make even more sense strategically. For the public, this would mean the best models in the world are about to become much more expensive to serve, and therefore much more expensive to use. That will put pressure on rate limits, pricing, and subscription plans that are already subsidized to some unknown degree. Instead of becoming too cheap to meter, frontier intelligence may be about to become too expensive for most of humanity to afford. Second-order effects; compute, memory, and energy are about to become much more important than they already are. In the blog they describe the new model as not just an improvement, but having 'dramatically higher scores' than Opus 4.6 in coding and reasoning, and as being 'far ahead' of any other current models. If this is the new reality, then scale is about to become king in a whole new way. It would also mean, as usual, that Jensen wins again.

English
0
0
3
105
R.
R.@RichDoesTech·
@Yuchenj_UW Been seeing everyone post these, honestly, I just think to myself "why"?
GIF
English
0
0
0
15
pc
pc@pcshipp·
Literally, SEO is getting more harder - 2 clicks - 11.8 average position - 327 total impressions Maybe SEO is 100x harder now
pc tweet media
English
80
2
113
14.8K
R.
R.@RichDoesTech·
Trying to publicly document the "messy middle". - Ready to launch on Android but Google play is approvals are being annoying. - Went back to focussing on Mac & Windows but had some latency regressions from feature creep (fixed now). - Need to test payments then good to go.
English
0
0
6
109