Copilot is the worst AI agent on the market. Over 50% of people I've spoken to that say AI is overhyped, when pressed, admit they were only using Copilot.
This would be profitable if they could ship garbage for cheap, a la Microsoft Teams or Internet Explorer. But Copilot is worse at integrating with Office than Claude!
This is because Copilot has aggressive context pruning to meet its price point of $20/month. That prevents the AI from meaningfully using tools or being multimodal or anything else their competitors have.
If they added a $200/month tier many of their issues would go away.
Their target is not coders, it is the professional world who do 90% of their work in Office applications, like me. A $200/m model absolutely does not fly when rolled out to entire corporations. It needs to be a $20/user/month product.
But I agree, it sucks. It is the only AI we are able to use at work and for tasks that it should be good at (compare comment sheets against a deliverable register and assign to specific packages) and it just can’t do it. It can read the spreadsheet and understand them just fine but outputs are garbled nonsense.
> A $200/m model absolutely does not fly when rolled out to entire corporations. It needs to be a $20/user/month product.
That isn't possible with the technology right now and it will not change. Multimodal computer use and long context for high quality outputs are expensive.
This is like complaining about buying an automobile because it's more expensive than a horse.
“Copilot” is not one product, it’s around 15 different products, seriously.
I think that people often compare apples to oranges by comparing the “copilot” they have in Windows/Office/Teams etc to Claude Code which is ridiculous.
A better product to compare Claude Code to would be “Github Copilot CLI”, but I haven’t seen the two seriously compared anywhere.
It’s odd to me how much Microsoft over committed to Copilot. They added unwanted Copilot buttons to laptops. They renamed Office to Copilot. And in the end, it’s a terrible product anyways. They can only get away with this because of the control they have over OEMs, distribution channels, and the inability for consumers to opt out of all of thus.
Meanwhile startups can’t compete fairly because they don’t have the same channels to flood with their own branding.
Copilot was by far the worst for coding. Not that the code snippets it would generate were not good, but due to the insane number of bugs in its UI. It would just spit out blank blocks thinking they contained code. When I asked it to repeat the steps which were empty, it would generate the same empty blocks like "here's your code" lol
We have copilot at work integrated very deeply in our E5 landscape. It def sucks in Office, and I can break Copilot very easily when building small notebooks. It often crashes and the next page build in a notebook doesn’t come close to the previous iteration. That’s maddening.
But, the Teams integration for meeting summaries and in-meeting “what did Bob just say about the data center project?” prompts is magical and very useful. If you live in meetings or are trying not to. They need to put that team on rescue duty.
I have no interest in Copilot from Microsoft in general but I do like GitHub Copilot overall. That said, I’m _very_ interested in a viable alternative. I only use it for “fancy autocomplete” and have zero use for the agent/chat capabilities (I use Claude Code for that). It’s been a year or so since I looked at and tried alternatives but when I last did, copilot was the best IMHO.
Microsoft is one of the least likely large companies to benefit from an AI boom. They don’t have the capacity to support OpenAI and their own foundational models, they aren’t providing a compelling story for wrapping OpenAI, windows continues to suck…
OpenAI signed an agreement with GCP , that should say a lot.
If one were to think a major stock market crash was coming up, led by the AI bubble bursting, but reinforced by the major self-own that is the Iran war, how would one best prepare ones investments?
Not investing advice, I’ve reallocated away from US domestic equities to international equities (VXUS) as a majority of a portfolio. This hedges against a correction from overweight Mag 7 exposure and US economic growth impairment from current policies (imho).
Either you need the cash now and are already in short-term treasuries or you're it it for the long-term and you'll be laughing at this question when the Dow hits 100k.
One should either weather out the storm or if one wants to cash out soon or manage their portfolio more closely they would pick the defensive assets they trust the most and hold until they stop thinking the stock market crash is coming up or stop trusting those assets. If they really think the crash is imminent, maybe investing some excess money into shorting the market while setting trailing stop loss would be a fun activity that might turn profitable
If you genuinely think these two events will shake the foundations of the US economy, then no investment is really going to protect you; you can't expect a system to shield you from the very thing that's breaking it.
There is crypto but even that got infiltrated by institutional wall street money. There are off-shore jurisdictions but the recent Iran war has showed these can be very vulnerable at a moment notice. There is China, but a Taiwan invasion could reduce your assets there to zero.
Honestly, I think the best bet is crypto/Bitcoin, by far. It operates across borders and still relatively insulated from government reach. Unlike gold, oil, or anything physical, it can be moved without physical visibility.
I don' think that logic is intended for the top... It is what you should do when you are closing to bottom or are recovering already but most of the market does not yet see it.
Being greedy at the top will take longest time to recover. Catching the falling knife.
The advice I have heard is if you think there will be a significant drop in the market you liquidate all your holdings while they are still high and then rebuy when the price is low. Granted this is a gamble though, if you’re wrong then you just sold all your stock and are no longer participating in the market plus you need to pay capital gains tax
spacex is particular is desperately searching for a bagholder. Leveraging the obvious synergies between rockets (a mediocre business), xai (a horrid business), and social media that lost more than half it's already modest revenues...
SpaceX has reasonable business in it. Not a hypergrowth one, but one which should be solid in long term.
To get that to work they just would need to discard Musk and most things with him. Stop trying to make starship a thing, dump everything attached to it. Make a long term plan to improve the core lift capacity with actually achievable improvements.
> To get that to work they just would need to discard Musk and most things with him
I'll have a pet unicorn shitting rainbows before Musk leaves one of his toys or we see in-orbit leading-edge (or anything close to it) processor production. SpaceX is a decent albeit capital intensive business if it's valued at $100-200B. At the proposed $1.5T+ valuation for this dog... the bagholder search is on.
60 comments
This would be profitable if they could ship garbage for cheap, a la Microsoft Teams or Internet Explorer. But Copilot is worse at integrating with Office than Claude!
This is because Copilot has aggressive context pruning to meet its price point of $20/month. That prevents the AI from meaningfully using tools or being multimodal or anything else their competitors have.
If they added a $200/month tier many of their issues would go away.
But I agree, it sucks. It is the only AI we are able to use at work and for tasks that it should be good at (compare comment sheets against a deliverable register and assign to specific packages) and it just can’t do it. It can read the spreadsheet and understand them just fine but outputs are garbled nonsense.
Copilot seems to hit the technical level I'm asking about much more reliably. It keeps a more grounded general semantic model.
Are you using the one that's part of Microsoft 365?
> A $200/m model absolutely does not fly when rolled out to entire corporations. It needs to be a $20/user/month product.
That isn't possible with the technology right now and it will not change. Multimodal computer use and long context for high quality outputs are expensive.
This is like complaining about buying an automobile because it's more expensive than a horse.
I think that people often compare apples to oranges by comparing the “copilot” they have in Windows/Office/Teams etc to Claude Code which is ridiculous.
A better product to compare Claude Code to would be “Github Copilot CLI”, but I haven’t seen the two seriously compared anywhere.
Meanwhile startups can’t compete fairly because they don’t have the same channels to flood with their own branding.
How TF can you go to market with such bugs.
But, the Teams integration for meeting summaries and in-meeting “what did Bob just say about the data center project?” prompts is magical and very useful. If you live in meetings or are trying not to. They need to put that team on rescue duty.
What are people using as an alternative?
OpenAI signed an agreement with GCP , that should say a lot.
The office workers who use their products do so grudgingly, since Office and Excel are so baked into business ecosystems.
Nobody will shed a tear when the aging dinosaur finally dies.
Unfortunately it has been given a transfusion and a new name, Microsoft 365 Copilot or some rubbish.
Not investing advice, I’ve reallocated away from US domestic equities to international equities (VXUS) as a majority of a portfolio. This hedges against a correction from overweight Mag 7 exposure and US economic growth impairment from current policies (imho).
https://www.axios.com/2026/03/27/stocks-trump-iran-nasdaq
https://totalrealreturns.com/n/VTI,VXUS?start=2025-01-20
https://www.apolloacademy.com/sp-500-concentration-approachi...
There is crypto but even that got infiltrated by institutional wall street money. There are off-shore jurisdictions but the recent Iran war has showed these can be very vulnerable at a moment notice. There is China, but a Taiwan invasion could reduce your assets there to zero.
Honestly, I think the best bet is crypto/Bitcoin, by far. It operates across borders and still relatively insulated from government reach. Unlike gold, oil, or anything physical, it can be moved without physical visibility.
Be fearful when others are greedy, and greedy when others are fearful. Etc.
Being greedy at the top will take longest time to recover. Catching the falling knife.
My feelings about these things don't come from markets.
You can't lose money sitting on cash. While when shorting your potential loss is infinite.
Most people are best off investing in index funds and forgetting about it for 10+ years.
and it is not good.
To get that to work they just would need to discard Musk and most things with him. Stop trying to make starship a thing, dump everything attached to it. Make a long term plan to improve the core lift capacity with actually achievable improvements.
> To get that to work they just would need to discard Musk and most things with him
I'll have a pet unicorn shitting rainbows before Musk leaves one of his toys or we see in-orbit leading-edge (or anything close to it) processor production. SpaceX is a decent albeit capital intensive business if it's valued at $100-200B. At the proposed $1.5T+ valuation for this dog... the bagholder search is on.