Yeah I mean this is closer to my use of LLMs, where I'm intentionally slowing things down enough to follow and course correct myself. This whole build the loop not the prompt idea seems to be advocating the opposite... I don't feel comfortable to not be the driver of the loop for a production system.
It's why I linked to a data driven study, and used language like "we had significant success with our loved one" and "if you want something that you can do to try".
Honestly this reads like an "aHCkTualLy!1!" from someone without experience of having a loved one suffering from a cancer diagnosis.
Perhaps you've yet to realize but shallow skepticism against every idea is also distinct from data.
I'm just struggling with this, surely you need inner depth knowledge to reason about the system and make some level of decision, at least around system design and architecture if not lower level implementation details? But it sounds like you're generating that knowledge each time through a system of agents? How do you have so much trust in a non-deterministic system, or are you deferring ALL decisions to these "loops"? What if you and a team member generate a dashboard and it gives different results because the agent(s) used a different methodology?
And surely cost plays a part here. This is giving you such productivity gains to boost revenue enough to outweigh what must be huge token costs?
As someone who makes a LOT of slides (I worked in consulting for many years, most recently as a manager at McKinsey), I literally tried every single AI slide maker on the market and this is actually game-changing software for me and many of my peers.
To the best of my knowledge, Brightdeck is the only software on the market that allows you to use AI to make non-trivial edits to existing PowerPoints. In my case an edit could be changing the format of an existing slide, incorporating a "sticky noted" comment on a slide , or taking raw notes I have on a slide from my deck's storyline (aka. ghost deck) and creating a slide from that which doesn't look out of place from the rest of the deck. In fact, I think they might also be the only software that allows you to use AI to add similar-format slides to existing PowerPoints while keeping the same layout/colors scheme...etc.
I have tried a few other tools and want to share my observations:
Gamma: Very easy to use and beautifully crafted. However, doesn't work with my use cases at all. If you try to upload an existing PowerPoint into Gamma, it destroys all your formatting and basically turns your deck into some simplified Notion file. However, if you don't care about preserving formatting, they're actually a nice option.
Genspark: I don't actually know how this company got to $200M ARR. Probably not from slides. At one point they (or their affiliate) pitched my team very heavily.
Claude: Claude Design can't make edits, their PPT plugin messes up even pretty simple edits if you're using it on an existing deck. If you're giving it a fresh prompt + fresh deck it actually works decently well (but it's super slow, like 20 minutes to generate 5 slides).
ChatGPT: ChatGPT.com can't edit decks (I believe; haven’t tried in a while), and the decks they generate from scratch are pretty basic. Worse than Claude. Their plugin for PowerPoint can only edit text--it can't change design/formatting.
Figma Slides: You can't edit decks. New decks from scratch are kind of nifty and have cool design elements though.
So...my conclusion is that so far Brightdeck is the only AI I can use to edit client decks. And the only thing I can use if I want to add new slides to a client deck. Congrats to the team on officially launching on HN. I'll keep paying for it and telling my friends as long as you don't take this comment as an indication to raise the price on me :)
Disclaimer: I found the product early through a mutual acquaintance and became an early user. Although I wasn't asked to write this, you should take this with a grain of salt, since I have quite a positive outlook on the company, their pace of releasing new features, listening to feedback...etc. I've been looking for this exact product for a while now.
Honestly the games very good, i played all the holes in one go. One thing I’d like though is for the ball to lose a bit less speed when it bounces off the wall.
I've been struggling with this exact problem lately and its been so frustrating. This looks like an awesome solution, I'll be trying it out later. Thanks for sharing!
Do you have any insights on how power was delivered to these circuits? Maybe it's done in the metal layers that were dissolved? Also, is it correct that there is no on die capacitance surrounding these circuits?
If anyone finds this thread because they or someone in their life is currently facing down a pancreatic cancer diagnosis I want you to know that we had significant success with our loved one by focusing, on our end, on diet.
The patient's metastasis markers were so high the value was literally off of the maximum value on the graph on the chart they gave us in the literature, and so, well beyond the level of being surgery eligible.
Over the 12 chemo cycles that number dropped to levels that cancer free people have, and they have gone on to outlive almost every statistic and remain cancer free to this day.
When researching pancreatic cancer following their diagnosis one thing that stood out to me is how the majority of scientific knowledge surrounding cancer addresses the cancer's metabolism. Pancreatic cancer is an IGF-1 (Insulin Growth Factor) metabolic cancer. This can be interpreted as the cancer uses sugar as its fuel source to grow, and in the absence of sugar can alter its internal metabolism to use an amino acid called glutamine as fuel instead. Glutamine is an amino acid found in animal products such as meat and dairy.
With this knowledge we went with a food regiment of removing ALL sugar, and animal products.
The results were significant. Even in their 70s they were able to do the full 12 cycle chemo treatment without needing to delay a single cycle due to negative health markers, and without any major side effects (except fatigue).
The tumor shrunk form 4.2 cm to 2 cm after 6 chemo treatments, and finally shrunk to 1 cm following their final treatment before surgery. (Compare this to studies on tumor shrinkage for the same cancer and chemo treatment, such as: https://www.healio.com/news/gastroenterology/20210722/early-... )
It is my opinion that at this time medical treatment is essential, both chemo and surgical intervention, but if you want something that you can do to try to increase the efficacy of those treatments I highly recommend this nutritional vector as well!
What are people doing at home? I have like 5 different apps I code on the $20/month Claude plan and like sure I can hit rate limits but - What are people doing to burn through $3k in tokens?
Mandatory voting is worthy of praise, some of the stuff you mentioned proves my point. Aboriginals are now part of that mandatory voting and as such the government bends over backwards to try and meet their needs while balancing them with the needs of others.
You are moving the goalposts, you asked to name a good government and I gave you one. On a spectrum from good to bad Australia is good.
If you ask to see a good horse and I show you one that can run at 30kmph, you can't in good faith complain that it requires oxygen and sometimes poops. It's not a bad horse because it can't fly.
Hi, I'm Richard. I am an IT professional with experience across multiple disciplines, particularly within the banking industry. Throughout my career, I have worked as a Full-Stack Developer, Solution Architect, Technical Business Analyst, and currently as a DevOps Engineer specializing in Middleware.
I am open to opportunities in any of these roles across various industries. My diverse background allows me to bridge technical implementation, business requirements, system architecture, and operational excellence.
Languages: Indonesian (Native), English (Professional), and Mandarin Chinese (Conversational, actively learning).
Feel free to reach out via email or LinkedIn if you would like to discuss opportunities or learn more about my background.
I think this is only going to become more relevant. I'm personally a $200/mo Claude Maxer and I know that the usage I'm getting on Opus 4.8 Max and (until they yoked it out from under me) Fable 5 is way, way more than what I'm paying them. At some point, this will turn usage-based and I will be hammered on it and probably forced to look at self-hosting. I think while the caps are there, even at $200, it's honestly not too bad if you're coding value into the market, but as soon as those caps come off for retail AI users, we're all going to have some tough choices to make.
Transmit and receive mechanical-television-style video through audio over a cable, WAV file, or pure software loopback. Classic 32-line club NBTV, and experimental wideband modes up to 480 lines that turn a 24-bit/192 kHz sound card into a surprisingly capable video link. It can even send files over the video channel as a stream of QR frames. In Python.
