ygouzerh

6 ms·

61.
by ygouzerh 7mo ago
That's quite interesting, well done! I haven't thought of this use case for embeddings. It open the door to quite many potential applications!
62.
by ygouzerh 7mo ago
Actually probably even cheaper, a generic scan to spot all the ships, and when it's done, just need to get images around the last location. Probably can use something like the Planet API
63.
by ygouzerh 8mo ago
I feel that it differs a lot between companies. It seems like corporate are having less an impact for now, as external innovation needs tailoring to adapt to its needs (e.g a security solution that needs 3 month projects to be tailored to t
64.
by ygouzerh 8mo ago
I am thinking to fine-tune it to recognize better my handwriting. It already works quite well by default, but my writing is just horrible, so it got trouble sometimes.
65.
by ygouzerh 8mo ago
It's wild, my Xiaomi phone charger came with 120W two years ago already, Apple seems so behind in comparison.
66.
by ygouzerh 8mo ago
The main root cause of the incident on their actions was actually due to Azure: https://www.githubstatus.com/incidents/xwn6hjps36ty points to https://azure.status.microsoft/en-us/status/histo
67.
by ygouzerh 9mo ago
It sounds like: Layoff --> increase short term valuation --> increase value per share --> owner of shares happy during buyback. After, it's true that having a lot of middle management can slow things down. On the other side, t
68.
by ygouzerh 9mo ago
From the company perspective, performing buyback when market is high is just throwing cash by the windows to over-priced shares. If they wanted to distribute cash, they could just use dividends
69.
by ygouzerh 9mo ago
It felt indeed that what the paper said is just: "If you are using a tool in order to make hard-work feel more easy... then your brain is not working as much"
70.
by ygouzerh 10mo ago
I feel the majority of junior job-hopping is due to the fact that they are often hired for really low, and then proposed just an incremental raise after two years. Instead, if they change company, then they got a big jump. At least, that&#x
71.
by ygouzerh 10mo ago
Thanks, I didn't knew about LiteLLM! OpenRouter have some interesting providers, like Cerebras, which delivers 2,300 token/s on gpt-oss
72.
by ygouzerh 10mo ago
The score are really, really close, it might be why
73.
by ygouzerh 10mo ago
I think the only reason is for legal purpose. If data is downloaded illegally from space, stored in space and model trained on it... it will be a mess juridically if someone complain. Same for model inference, it will be hard for a governme
74.
by ygouzerh 11mo ago
It's nice as well for location that are banned to use private US models. Like here in Hong Kong, Google doesn't allow us to subscribe to Gemini Pro. (Same for OpenAI and Claude too actually).
75.
by ygouzerh 11mo ago
I feel the issue is that now, what would be done by entire teams (networking team, storage team, database team), is now perform by only the same DevOps team. We have way less time unfortunately to dig into each tech, business is pressing us
76.
by ygouzerh 11mo ago
It's mostly for AWS context, when you want to host for example your servers in a private subnet, but you want to allow access of small part: you can leverage a NAT Gateway to be the public entrypoint + some security groups as gatekeepe
77.
by ygouzerh 11mo ago
I am curious: why would you you like to have that? (Genuine question, I am personally so scared about the AI going crazy and putting slop everywhere that I often ask it to focus on a single well defined area first)
78.
by ygouzerh 11mo ago
Good idea! I found Gemini have horribly slow for anything
79.
by ygouzerh 11mo ago
GPT-5 was horrible. It produced AI slop have immense speed, which is quite tough when other coworkers ask to review their PR...
80.
by ygouzerh 11mo ago
Definetly! I put in the instructions.md file to check if the code is well conform to the latest doc using Context7, works quite well!
81.
by ygouzerh 1y ago
Nice, seems way cheaper to use than the new Cloudflare Sandbox SDK solution
82.
by ygouzerh 1y ago
If they are spending their time fiddling around disk resizing and package upgrades, it's time they are not spending helping deliver real business features. This part can be outsourced to a PaaS company, so that the company engineers ca
83.
by ygouzerh 1y ago
Interesting read! It kind of a became a rule nowadays to never store logic in the database nowadays, but it's always refreshing to see when it could work. Please give us an update in few months of how it is going on the long run!
84.
by ygouzerh 1y ago
One thing however to note, is that by having a different non-prod and prod environment, it will be possible to test only the application, and not the infra. Which means, that if they want to test what it will look like running in cloud for
85.
by ygouzerh 1y ago
Yes, everyone forget to compute man-days in the cost calculation
86.
by ygouzerh 1y ago
I will recommend to try GCP or Azure, the complexity is lower there! AWS is great for big corporate that needs a lot of lego pieces to do their own custom setup. At the contrario, GCP and Azure solutions are often more bundled.
87.
by ygouzerh 1y ago
Single server is very cheap for hobbyist. Just something to consider if you are in a professional environment before switching your entire infra: maintenance cost is expensive. I strongly suggest to throw man-days in your cost calculation.
88.
by ygouzerh 1y ago
Github Copilot could help you, you can switch model from different providers on the fly (supports Anthropic, OpenAI, Grok,...)
89.
by ygouzerh 1y ago
Boston Dynamics is very good at the mechanical part of robotics. They only start recently to take a look at integrating more autonomy and long term decision making.
90.
by ygouzerh 1y ago
From what I understand, WASM is purely functional, to perform side effects, you will need a runtime like WASI to implement it

More ›