7213Ejaaz

wow. Apple's suing openai for stealing hardware secrets for their upcoming Ai devices releasing later this year.
apparently 1 employee convinced apple team members to bring proprietary hardware devices to interviews.
Openai and apple are going head to head on ai consumer devices later this year with apple planning a new airpod model with cameras, smart glasses and a pendant device.

this week's been a total win for ai, so many new models and features to try out this weekend:
> 1 new frontier model: gpt 5.6 Sol
> 2 new crazy cheap models that are 80% frontier: grok 4.5 and meta muse spark 1.1
> 2 new image models: muse image and seedream 5.0
> prime intellect raising $130M to push out more open models
and google's releasing 3.5 next week, deepseek v4 too, grok 4-5T model in a month and GPT 6 in a month
its all happening so fast

1 of the biggest shifts in ai has been training AI chips to inference-optimised chips - and for GOOD reason:
inference is where the money (and intelligence) is made
> anthropic is rumored to make 80% revenue margins on inference. they’ll be profitable before any other ai lab.
> apple, meta, openai, google are all making their own custom ai chips for and inference
> chinese model labs (despite the hardware handicap) build frontier intelligence purely by forcing the model to think of longer.
point is: inference is going to out-grow another ai service for a country mile.

as america turns 250 it’s now more important than ever to remember that this country’s secret sauce is its people and their relentless pursuit of bold, ambitious ideas. that’s what’s kept america winning and what will keep them winning going forwards.
looming threats of foreign competition across AI - stealing model designs, natural resource restrictions, scaling wars etc will increasingly become a problem
we need to protect our data centers, people but most importantly our ideas. that’s what keeps us at the frontier across *any technology* over the past 250 years
you can’t outcompete genuin

it looks like openai’s betting big on inference to beat anthropic over the next 6 months:
> between their cerebras and upcoming jalapeño ai chips, openai will have the ultimate hardware stack optimized for GPT
> they’ve cut inference costs by 50% (per the information) which will save them $10Ms
> gpt 5.6 leans heavily on agentic reasoning and execution (inference heavy)
inference makes up 40-50% of compute expenditure at labs, slashing that in half frees up compute to use for training - the compounded effects of this can’t be understated imo
inference is also a huge money maker: anthropic’s r

i don’t buy the meta and ai memory fud:
- google just shut off meta for using too much compute, why would they then need to sell excess compute?
- meta owns one of the largest fleets of gpus, they’re likely selling the older ones for inference (like spacex) while they utilise newer rubin’s.
- memory is without doubt the scarcest commodity in AI, demand outpaces any projected future supply by 5-10X.
- in the event someone designs a chip, model or agent that requires less memory - this will only result in MORE memory demand
have you been watching how much memory these agents gobble up per sess

so it looks like there’s a few key trends converging in ai recently
- frontier models now pose a legitimate threat to government security calling for government intervention (rip fable)
- this means future model releases will be private (restricted). any future public releases will almost certainly be handicapped
- at the same time, chinese open models have caught up to frontier capability. in <6 months they will be mythos-level
- open models are cheaper, can be run privately, hard for governments to regulate
- let ai talent is consolidating to the leaders. anthropic and openai are absorbing

the problem is even if anthropic magically creates a hack-proof safeguard for mythos - there will be other attack-surfaces for actors to exploit
cloud services, employees, data centers, prompt-injecting
there can never be a 100% fool-proof model, we’ll just have to get better at defending like we did if every other software upgrade

way i see it there are 4 companies with a shot at owning 70%+ of model market share
anthropic, openai, spaceX and… open source
- anthropic & openai are on the verge of self-improving models. end game.
- but spacex acquired cursor in the final inning. their harness gives them a legitimate shot at catching claude and gpt in coding. IF elon pulls that off then he’ll win on raw compute. no one has more gpus than him
- open source wins if frontier models price out regular consumers. they’re good enough for 80% of uses. we’re starting to see this with microsoft switching to deepseek
each of them are

AI is getting way too expensive and its the reason markets might blow up BUT not for lack of demand - its the exact opposite:
> demand for ai is skyrocketing. the models actually work but the tokens are too expensive.
i switched to claude fable 5 this week and hit usage limits almost immediately
who wins? chinese open models. fraction of the size, 80-90% of frontier capability, costs 1/200th of claude fable, gpt 5.5 high.
ignore the benchmarks btw, no one in the real world actually cares, its the bill they have to pay:
> the number of US startups shifting their claude subscriptions to cheape

memory and ai stocks cratering today is laughable tbh.
anyone taking a long-ish view knows we’ll need so much more of this stuff
gpus, consumer ai devices not to mention the shit ton of memory needed to maintain everyone’s chat history and context
i’m personally running out of memory slots across claude and gpt subs (max tiers), running multiple agents (honestly because i can and it’s helpful with menial tasks) and i know im not the only one.

hot take but i just don’t think we’re in an ai bubble. ai capex isn’t levered and most importantly we are limited by the literal laws of physics in terms of scaling:
> you can raise all the capital in the world ai capex but you can’t spend it. construction is slow, regulations are blockers, literal power grids can’t keep up.
all of this needs to be unblocked, that will take time.
- google’s $80B equity raise is just creative financial book-keeping. 30B is used to offset taxes, $40B vs the $180B they committed is not a levered position. they have $126B on the balance sheet.
- sk hynix announc