people always ask what I do or how I got into AI… my background - over the past 6 years, have built AI systems for startups, mid-sized companies and global enterprises - youngest to get a phd in biostats from uc san diego - previously led AI/ML research initiatives at Harvard - recruited by the CIA at 18 and yeah on paper it seems like im this 200 IQ mf but funny enough I actually got rejected by pretty much ever college out of high school... and was a straight dumbass growing up (still am lol) but from that moment on I told myself theres no way in fuck ima be a bum… and just went head deep and pretty much taught myself how to code and theoretical math (which is what every AI model is built off of) and shit didn’t really come naturally to me either honestly... it really was just a fuckton of time spent throwing shit at the wall and slowly becoming less and less ass until one day I somehow became kinda decent at this stuff haha then out of nowhere through a professor at school, I got in contact with the CIA and later Harvard and ngl all that put my life into a whole different trajectory that I never envisioned for myself (note I never worked for the CIA thank god in hindsight getting a security clearance is prob one of the most painful processes in the world but was able to leverage that to get other opportunities like working at harvard and shit)

a close friend of mine is a software engineer making $350k+ a year and he finally got hooked on AI this week bro tried claude for the first time and knocked out all his work for the week in a day and was obviously amazed haha his exact words were "my job has become just waiting for claude to finish now" and the crazy part is most of his coworkers don't even use their github copilot credits besides maybe for unit tests and that's it on X I think we forget we're in a tiny bubble where everyone has used this stuff daily when there’s ton of people out in the real world (whose literal job is deploying AI agents lol) have barely touched this shit

apple never sold the ipod by bragging about storage specs they sold the idea of "1,000 songs in your pocket" I learned this exact lesson when getting customers my software when we first launched it was costing us ~$350 to acquire a customer on meta and the product was great. anyone who watched the demo loved it the problem was the headline on our website was all about the tech: "powered by AI agents trained on X, Y and Z" and after talking to customers I realized that line was pushing people away bc to most people "AI agents" sounds like work to them they're thinking: now I gotta manage these things? learn how to prompt? figure out which model to use? fck that they don't care about any of that. they just want the outcome with the least amount of effort possible so I deleted the AI agent stuff and the subheadline just became "get X without Y" meaning all you have to do is click a button and get Z press a button, get the result This one change dropped our CPA on Meta by 70% it was unreal same reason why you see a software like ListKit for cold email absolutely ripping if you've ever run cold email campaigns, handling all the domains is horrendous to do yourself and they handle all that for you just by clicking a few buttons it's a lesson I always have to relearn just sell the mf benefits and not the features

I have a free lifetime GPT Pro account and I still pay for Claude Code every month that should tell you how much more I like Claude Code lol here's my current AI stack that I use day to day: claude code: what I use for basically all my AI coding. I do have codex set up as a skill to give a second opinion for more complex stuff if/when I need it hermes: this is my daily driver. I barely even log onto chatgpt or claude anymore. they're connected to my hermes and I just talk to them through hermes because I keep all my skills on here and its connected to my higgsfield acc and my apis I use wispr flow: I pay for this bc I'd rather just talk to my agents than type everything out grok bot: I post on X so this is my live line to what's happening on here. it's connected to my X account so it can find trends and mine data on X super easily. also great if you're a beginner and not very technical compared to hermes obsidian: honestly just a swipe file for ads and posts I like I know a lot of people use obsidian as a second brain but it's never been that useful for me personally maybe just not my style (although though I've built them before for clients and they all love it)

went out to canada last month to golf at the fairemont in banff - one of my bucket list courses actually some of the most unreal views I have ever seen

took a long break from posting bc I made it my goal this year to actually get good at running ads and more importantly converting cold traffic so earlier this year I said fck it… ima try selling an AI software with meta ads and to no suprise shit was tough af haha because it’s not just making good ads when selling a software… once they’re on your page you still gotta get these people who know nothing about you to pull out their card and actually like buy from you haha (it didn’t help I picked a giga saturated niche to sell to as well lol) along with that getting the actual economics to work out so you’re actually profitable and building out a good product too of course so they actually stay it didn’t happen overnight whatsoever as this required tons of testing with the ads, messaging, LP, VSL, etc but as of right now we have our $67 front end offer converting at 5-7% on purely cold traffic with a CPA of little over $100 and with stacking on upsells we got our AOV to $229 I'm by no means an ads expert nor do I plan to go all in on this SaaS as my AI consulting agency has and will continue to be my main focus but it’s been a sick journey these last 5ish months seeing it come together

if you’re looking for an entry level job as an engineer in tech, apply for jobs through personalized outreach on Linkedin or cold email literally all my friends and family do this and it prints interviews… - pick a company you want to work at - build a project relevant to what you would do if you actually worked there & match the UI to the company’s colors - send the deployed link along with the Github repo directly to the hiring manager through email and LinkedIn with claude code or codex you can spin these up with like no work… just make a skill where you give it a company website and it deploys a project relevant to the company 99% of people aren’t even going to look at the code anyways... they’ll just be impressed you went out of your way to do all this and now you resume isn't giga buried under 1000s of applications they get Linkedin/Indeed

a huge misconception I see with people selling AI software especially from engineers/devs is they love to lead with how amazing their tech is but im telling you there is a large piece of the market (esp outside of twitter) that doesn’t give a fuck about any of that… these people they understand AI is important and want to use it within their business BUT they have absolutely zero desire to learn it… from code, learning new tools, fuck even prompting… they don’t want to touch any of that and to a lot of people when they hear shit like "you’re going to all these AI agents, claude skills, APIs, MCPs, etc…" to them it often just sounds like a lot of work for them when the better route is often just positioning the software as dumbass friendly like where they get their desired output with as minimal effort possible

apparently twitter algo is not wack af anymore thus I shall start my return with a quick video on the standard process/mental framework I use to evaluate any AI systems you deploy that way you can easily identify if the AI agents you deploy are actually doing the correct job and more importantly so that when you make changes to the agents you actually can identify if you are improving or not

obviously with the rise of AI, it has become 100x for the everyday person to push out content but something important to keep in mind is the difference between content that’s like wow that’s so real I can’t believe AI made that vs content that actually converts and makes money bc the biggest trap people are falling into it seems like with AI content is that yeah it may look good and you cant even tell it’s AI but a video looking realistic obviously doesn't determine if an ad will print or not as there are obviously tons of ads made by human people that are 100% real that dont make shit and this isn't me hating on AI ads as I literally run a ton of them myself but it's also important to understand there exist much more important levers to pull such as the messaging/angle, emotional triggers, etc so essentially just DR/marketing fundamentals

this is actually so sick especially the fact it allows for mixed-inputs meaning you can embed video + text together and get a singular embedding out a good use case for this would be building out an AI swipe file for ads like for a specific ad you can take the creative, script, hook type, angle, stage of awareness, USP, whatever you want really and create a singular embedding out of them using this new Gemini model and makes it 100x easier now do searches like "I wanna see supplement ads that focus on gut health" as the search will be done on the text layer and the creative layer now where as before it could be only done via the text so if you wanted search on the creative layer you'd probably would've had to get an analysis of the actual creative via GPT or Gemini to convert into a text based explanation and then append everything all together

if you want the highest quality AI UGC or talking head ads, Sora 2 + Kling Motion Control is probably the best combo... as they are extremely complementary for each other where Sora struggles is they have this distinct plastic/mannequin look to them where you can just tell its made from Sora and you don't really get true character consistency either bc they don’t allow photos of people to be used in their reference pics but in terms of natural movements and pacing which is extremely important obviously, they're by far the best so use Sora 2 to create your base video so that you get the natural movements and the pacing of the character… then throw that into Kling Motion Control where you use a realistic looking reference photo from Nano Banana Pro and now you get the benefits of Sora’s natural movements and pacing... but now it also looks more realistic and you also get true character consistency to create long form videos as Kling allows for reference photos

if you're making any AI UGC, all video generation models have like this distinct static esque voice to them I'm not sure how to articulate it tbh especially if you watch it on an mobile it's usually more apparent there compared to watching it on my mac regardless, would highly recommend throwing any AI videos through Adobe Podcast's Enhance Speech at the bare minimum if you're not using any other voice changer as it's super easy to use... like it's just drag & drop and then the tool will just clean up the audio from there

read this shit if you’re doing anything AI related

random observation as for whatever reason much of my work lately has involved image analysis of some sorts with LLMs but despite GPT being ass af lately... when it comes to extracting visual details from images into a summary or a json, it has been absolutely shitting on every other model

when it comes LLM benchmarks analyses I rarely ever look at them bc at the end of the day the best models are the ones that YOU like the most there is simply no benchmark that can capture this bc everyone structures their prompts and uses LLMs in so many different ways and plus everyone has so many different and nuanced use cases with AI (even if you subset to something as simple just plain writing) that it's impossible for a benchmark to a make blanket statement on which model is superior a great example of this is I actually prefer Gemini for writing ad scripts especially if I want the best result in one shot and I have no reason why other than I just simply like the outputs better and again this isn’t even me saying Gemini is the best LLM for writing ad scripts... it’s simply my choice of model for creating ad scripts that seem the best to ME based off the way MY PROMPTS are written and structured again everyone has different preferences in what a good ad script even looks like and everyone has different ways in which they provide context and different ways they write and structure prompts the key thing to understand is you only come to these conclusions on which model feels best to you not by making assumptions based off what other people say or what benchmarks… you come to these conclusions from actually testing shit out yourself and seeing whats good vs whats not good

had to go on a small break from posting was just locked in af other projects and honestly kinda burnt out a bit from posting I think from just relentlessly only tweeting about strictly AI for months I just got kinda bored of it and also kinda forgot that this supposed to be a PERSONAL brand not just an AI account especially considering I don’t really shill anything on here and I have many other interests business related as well outside of just AI regardless still crazy af to me that 10k+ people actually follow me and even remotely give a fuck about what I have to say but my break is over now and my creativity and desire to write shit has returned I will be posting more consistently (but no where near at the rate I was doing before) luv u all