Baffles me that 18 months ago we were discussing a “VC winter” and possibility of a ‘07-like recession to reign in the excess of late-2020 & 2021.
Now it feels like froth and companies that barely do what they claim is everywhere. Yet so much VC money is still on the sidelines.
People say there's more AI investment, but AI has always been heavily overinvested since 2015. Even back then, fresh PhDs could expect million dollar annual TC. People who believed AI would eat the world have been buying Google, Meta, Tesla stocks for years because those were the closest. Those stocks aren't rising much. Only recently did retail investors realize that maybe NVIDIA would be profiting from the AI boom as well.
You have to take into account all the delays. Meta didn't drop in stock because they were planning for the Metaverse, it dropped because they admitted that Facebook was no longer their core business. Everyone knew FB was dead for a while, the name change was a catalyst. The metaverse isn't terrible business though - Roblox is bigger than many banks.
There is really nothing to validate this statement. Roblox is a $20B company. Facebook is 60x bigger. Shareholders did not believe that the metaverse was a large enough opportunity for Facebook to go "all in" on.
But is the metaverse worth over a trillion dollars? If it's not, then what's to say social media is?
Then there is the AI hype. You see someone like rabbit selling a lot and getting more money... and that's turning into another Jucero.
And that makes OpenAI look a LOT like Theranos... all hype and no substance. I have a pretty solid argument that this is likely the case. Lets go back to GPT 3... Sam in front of congress, where all talking about the singularity. The deal with M$.
That MS deal had ONE interesting bargain in it. They could have everything till OpenAI got to AGI. Thats a pretty drastic bet to make.
And here we are, after board fiasco, where any one who might have said the word safety 12 months ago has quit.
It's the google engineer who thought their efforts were AGI, it's the MS paper saying hints of AGI. I think everyone at open AI got high on their own supply. I think they had an idea that if they just fed it enough info, if they gave it EVERYTHING that some how real AGI, or sentience or sapience would emerge.
Candidly from where they sat it probably looked pretty close, and it's hard not to have rose colored glasses. More so with Sam as the hype man on TV. The problem is that it completely ignores the history of every AI effort that has gone on since the 50's ... where they think they are close and then reality swiftly kicks the project in the balls. It's not like this very thing hasn't happened every time before. Us getting close to AGI is like fusion being 10 years or 25 years away, always... you wont ever get there if you keep covering half the distance to the goal.
Meanwhile google, FB, M$ cut staff and bought hardware on this bet. The thing is no one has ANYTHING good to show for it. And for as much promise as it has, I haven't seen a "reduces hallucinations" paper, or one that has an inkling of how to make the systems more honest.
My bet, that there are a bunch of engineers out there working on cool shit. They are bootstrapping because they dont need VC's any more. There isnt a massive investment upfront in "infrastructure" and most of them are skilled enough to tie scale to revenue and do it so there is margin. I suspect that you won't get another google or FB out of this one... you're going to get 1000 companies that do 5-10 million a year with sub 10 total staff.
While conversational AI is nowhere near as good as people are made to think, there are two places where it has really shown itself to be good:
1. Writing the first draft of things. Be it code or a email.
2. Turning unstructured data into structured data
By and large, those use cases aren’t very sexy, but the have the potential to accelerate the speed at which many things are done. And with less manpower.
Talking this through with you, I’m starting to think that two things are going to happen:
1. A significant number of people who are or will be laid off in an effort to reach profitability are going to either found new companies or take senior roles on newly founded ones.
2. A mild recession will occur for some reason (be it layoffs, consumer debt, political unrest, a long tail event, etc). That’ll finally lead the fed to lower interest rates, which in turn starts the process of a true VC driven boom cycle.
Reinforces my suspicion that we are very early on in a super-cycle.
Datadog sold what startups wanted. You want an observability platform, they would sell it to you piece-by-piece. This was great to the higher-ups; in an era when data science and dashboards were all the craze, every startup you went to had a monthly Datadog bill in the thousands. They were hooked.
Then the economy went to shit, and you know what everyone realized? Datadog was useless. They were taking our money hand-over-fist to give us features that didn't matter and detracted from developer funding. When the higher-ups decided to stop trend-chasing, they made the call to drop Datadog and usually replace it with something else (some lumbering Prometheus stack).
The moral of the story isn't that you can see a need and fill a need; the moral is that people will abandon half-measure solutions. OpenAI isn't filling a need, they aren't directly addressing a problem people have been demanding a solution for. They're selling a trend that you can join-in on for a per-call fee from their API. Much like Datadog, one day startups will start removing ChatGPT to replace it with local LLMs, and then probably remove LLMs altogether later down the line. Mark my words, the most successful businesses that come out of the next 10 years will barely even mention AI as a footnote in their product description.
Video games having rich NPC dialog (so many memes are gonna come from this). An agent that can have some sense of prioritizing email/text/chat to alert you. Someone to proof your emails. DOCUMENTATION has been the bane of every dev's existence, LLM's are useful here.
LLM's aren't fire, they arent steel. I suspect that they are going to fit into our lives much the same way "maps on cell phones" did, and just make a lot of stuff smoother.
I think LLMs will be the robocalling of the 2020s. You'll stand in the self-checkout at the pharmacy, waiting 5 seconds for an LLM to generate a unique, awkwardly-worded way of asking you to clear the bagging area. This is what we're damned to, not necessarily because AI is bad but because executives don't know what to use it for.
You are 100% right with this conclusion, and I think you said it better here than in my post.
They are great at some tasks, but your hand holding it through them, and checking its results. Its a boon to productivity (and for me that was writing docs, and cleaning up email)...
I use it more like an advisor that I do not trust. Its like I asked the question on reddit and picked the top answer... It might be spot on, or it could be dead wrong, or it could be trolling me. This is very useful but one has to be aware of its shortcomings.
How do you get people using something that is helpful but has a large list of caveats?
https://investors.datadoghq.com/news-releases/news-release-d...
> That’ll finally lead the fed to lower interest rates,
Interest rates are low. Having them lower is DUMB and we saw what happened. The fed likely wants things to stay put incase inflation creeps back and then crank them up again... Not going to happen till well after we see a new president and maybe not even after that.
We dont need the rates to change to see the market move with VC money. Look at what they were in 1999 compared to now. The money was flowing then, it will flow again regardless of rates. VC's are just going to have to make some better bets.
Does it do that without hallucinating? I see lots of people talking about this, but not mentioning quality.
I didn’t think it was a fair comparison at first either, but based on this standard it actually might be. Theranos absolutely made something. It had a device that took blood and gave back results. The fraud was that the results were “imperfect”.
Overhyped? Sure. Compare them to TSLA at least. But they quite literally changed the world by mainstreaming a single tool, and got entire continents suddenly making a sector a priority by sheer existence. Very few companies can claim this - NONE other than OpenAI can claim it in such a short lifetime.
But makes me think about how things like social media were supposed to lead to various positive outcomes that they certainly haven't.
Immigrants are the canary in the coal mine. When more people are unemployed, the local talent is already fighting for available jobs, so companies don't hire as much from abroad.
This is based on relatively flimsy evidence, but it's how I currently see things.
When finance starts to lay off lots of people people, it's because people are spending less. When people are spending less, it causes recessions.
A very strong metric to watch is interest rates. Most recessions hit when interest rates are 0%. Rates are lowered to encourage growth, but when they're at 0%, the lever is stuck.
Starting to think there won’t be any big broad recession similar to 2007, but more of a rolling one bouncing from sector to sector. Which is kind of how it should be.
In some ways I think tech is reaching the end of its recession. Where it goes next will be interesting.
Despite my personal unemployment it seems things are better than they were in early 2022
Crypto really poisoned the well, IMHO, bringing outright scams front and center. The dotcom bullshit from 30 years ago were mostly decent ideas executed too early. Speculation on electronic baseball cards depicting cartoon monkeys was always madness.
Based on your references, I assume you were around for that cycle -- in hindsight, what did/should you have done during that stage of the boom cycle to prepare yourself to thrive during the inevitable bust?
There are obvious answers like "live within your means, save money, and invest conservatively." But curious about deeper ones. Particularly when it comes to career trajectory.
LLM capability scales with money, so deep pockets will win, that simplifies things a great deal.
Microsoft looks like a better place to work these days strategically. They were an early Facebook investor and, as evidenced by the Sam Altman incident, essentially own OpenAI. They have now shown a pattern of being early to see disruption and get their money and influence in. They are diversified and don't make big money on search, which seems like the obvious area of disruption for AI. I think the rise of LLM helps them more than any other company, but we'll see?
Get the lawn mower running: https://chatgpt.com/share/1ec04770-dbe3-4904-96bd-e89cd65893...
Can't help with the beer though, try DoorDash.
Anyway every company stole the NLP/LLM/Rag model out of in chat bots got better results no need for ui budget (looking at you salesforce). Some tasks like call centers were pretty much also made for LLM. Other AI advances like sorting packages and treading the line between algorithm and AI do well.
More importantly big companies are making tons of money and there’s not been appreciable advances in technology. LLMs will be local and integrated it won’t be called AI. Workflows will improve etc. right now it comes from big tech with cash on hand it looks like big iron from the 80s. The internet didn’t see anyone control it and was relatively cheap to get into. Create a 3d image scanner that can process millions of say walnuts on an assembly line as good or bad isn’t sexy and niche… and expensive.
So we are in a lull companies are bringing dev in house and aren’t being experimental. Unless you have one of 7 big tech firms behind you won’t get looked at. So right now it is get to the state in b2b to just get bought.
Also b2c is controlled underneath by AWS or a big cloud provider.
Good thing companies are realizing they don’t need Google levels of high reliability. I see a real progression back to on prem with slimmed down K8S variants with Chromebook like devices. Google spent a lot making those “just work” as anyone installing even Ubuntu can tell you. But we are basically at the point someone will create a business like Steak deck with hardware compliance. Servers where apps can be deployed like VDIs and a resurgence away back to guy in his garage software development. It’ll still need VC funding but with secure infra you can have on prem SaaS catering to the mid market.